mmmmn

0606月25，26，27，28，29，30日

mmmmn — Sun, 20 Aug 2006 23:59:02 +0800

1、Re: how to start portmapper service in Solaris 9?

--------------------------------------------------------------------------------

From: victorfeng1973@xxxxxxxxx
Date: 30 Jun 2006 13:18:40 -0700

--------------------------------------------------------------------------------
The rpcbind is running. Thank you, Casper!

Victor

Casper H.S. *** wrote:

victorfeng1973@xxxxxxxxx writes:

how to start portmapper service in Solaris 9?

It should be running by default and is really hard to get rid of.

Look for the "rpcbind" process.

Casper

--------------------------------------------------------------------------------

References:
how to start portmapper service in Solaris 9?
From: victorfeng1973
Re: how to start portmapper service in Solaris 9?
From: Casper H . S . ***
Prev by Date: error when installing Flash Archive.
Next by Date: Re: binary emulator for sparc cpu on x86 systems
Previous by thread: Re: how to start portmapper service in Solaris 9?
Next by thread: System file page cache - read-ahead
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(07)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

2、Re: tail memory allocation

--------------------------------------------------------------------------------

From: Paul Pluzhnikov <ppluzhnikov-nsp@xxxxxxxxxxx>
Date: Thu, 29 Jun 2006 08:59:39 -0700

--------------------------------------------------------------------------------
"tmp123" <tmp123@xxxxxxxxx> writes:

foo | tail -1000 > /tmp/log.txt

However, there are the doubt about how many memory "tail" uses.

In the time it took you to post this message, you could have
performed a trivial experiment and answered your question yourself.

dd if=/dev/zero | od -cv | tail -1000

In another window observe the 'tail' process size with 'ps', or
'top' ...

It stores internally all the received data ...

Any tail that did this would be quite unusable on large files.

Cheers,
--
In order to understand recursion you must first understand recursion.
Remove /-nsp/ for email.
.

--------------------------------------------------------------------------------

References:
tail memory allocation
From: tmp123
Prev by Date: Re: Suppported disk drives
Next by Date: Re: x86 Solaris 10 (10 slices or 16 ?)
Previous by thread: tail memory allocation
Next by thread: x86 Solaris 10 (10 slices or 16 ?)
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: CAS, Lambda Calculus, Continuation and Functional Programming
... >>Recursion that's equivalent to loops can always be optimized back into ...
and that tail calls are equivalent to loops. ... If a loop is formulated in a recursive
way then ... >> that makes tail call elimination less attractive for compiler writers.
... (sci.math.symbolic)
Re: off-topic: Why is lisp so weird?
... >> recursion that is used to express an iterative process, ... That's
not tail _recursion_. ... >> not anything unique to Lisp. ... The same
happens in C. TCO has nothing to do with modules. ... (comp.lang.lisp)
Re: whose 8051 cc overlays static inline stack frames
... > optimisation, check compiler writer's literature for 'tail ... > If
there are separate stack frames for the functions, ... > advantage of the tail recursion
gets questionable due to the ... (comp.arch.embedded)
Re: Proper tail recursion
... > Instead of doing tail recursion elimination, ... > general tail
call optimization. ... > the current stack frame before calling spam. ... >>
the current executing function or generator object. ... (comp.lang.python)
Re: newbie question : summing elements in a list
... >> place that recursion is required nor even always encouraged. ... >>
Recursive solutions can use stack and you can run out of stack. ... >> that the Scheme
language defines that tail recursion will not create ... > ONLY FOR TAIL CALLS!
... (comp.lang.lisp)

--------------------------------------------------------------------------------

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

随机文章：

0605月19，20，21，21，23，24日 2006-07-26

0604月19，20，21，22，23，24日 2006-06-26

0604月10，11，12，13，14，15日 2006-06-15

0510月25，26，27日 2005-12-09

0508月7，8，9日 2005-09-13

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0606月7，8，9，10，11，12日

mmmmn — Sun, 30 Jul 2006 21:50:40 +0800

1、Re: iostat -- is there something better?

-----------------------------------------------------------------------------

---

From: Robert Lawhead <news0000.5.unixguru@xxxxxxxxxxxxxxx>
Date: Tue, 13 Jun 2006 10:18:50 -0700

-----------------------------------------------------------------------------

---
Jay G. Scott wrote:

i've read iostat's man page, maybe that's what i did wrong.
if i do:
iostat disklist -l 20 whatever
it
1. doesn't give me everything in my disklist

Can you provide an example of the arguments you used and
output from iostat? Did the switches match the names?
(For example if your list contained c0t0d0s0, did you use -n).

2. still cuts the list off at 20

Isn't that why you used "-l 20"?

3, lists disks i don't care about.

depending on options, the disk goes on the left or the right....

And this matter why?
I think "-n" (names on right) is newer option, so perhaps
behavior is rooted in backwards compatibility.

w/o -l 20 i get far more output than i can digest. also, i can't get
things labelled by mount point.

Works for me, See below.
I think mountpoints are only given with "-n" output,
which seems sensible because names printed w/o -n don't
resemble anything from the output of "mount" or "df".

the thing just feels chaotic. is there something better?

For what specific purpose?

One thing that I think is a bug is iostat's interaction with /dev/fd.
For example I think if I have a file called 'disklist', that contains
the names of the devices to list, then

$ iostat -nxpm -l 10 <(cat disklist)

should work, and doesn't.

% iostat -nxpm # line numbers added by hand

extended device statistics
r/s w/s kr/s kw/s wait actv wsvc asvc %w %b device
0.8 0.3 18.9 5.2 0.0 0.0 9.7 4.5 0 0 c0t1d0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t1d0s0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t1d0s1
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t1d0s2
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t1d0s3
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t1d0s4
0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.6 0 0 c0t1d0s5 (/export/stuff)
0.1 0.2 5.2 2.5 0.0 0.0 0.5 3.6 0 0 c0t1d0s6 (/export/home)
0.7 0.1 13.6 2.7 0.0 0.0 14.5 4.9 0 0 c0t1d0s7 (/usr/local)
0.2 0.1 8.5 1.3 0.0 0.0 0.0 7.4 0 0 c1t1d0s0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c1t1d0s7
1.8 0.9 23.3 21.8 0.1 0.0 42.5 11.4 1 2 c0t0d0
1.3 0.4 16.5 0.4 0.0 0.0 13.2 11.0 0 1 c0t0d0s0 (/)
0.0 0.0 0.0 0.0 0.0 0.0 22.5 24.9 0 0 c0t0d0s1
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t0d0s2
0.4 0.5 5.2 19.8 0.1 0.0 91.1 11.8 0 1 c0t0d0s3 (/var)
0.0 0.0 1.6 1.6 0.0 0.0 135.9 16.0 0 0 c0t0d0s4 (/opt)
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 fd0
0.2 0.1 8.6 1.3 0.0 0.0 0.0 6.3 0 0 c1t1d0
0.2 0.1 10.3 1.3 0.0 0.0 0.0 8.4 0 0 c1t2d0
0.2 0.1 10.3 1.3 0.0 0.0 0.0 9.7 0 0 c1t2d0s0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c1t2d0s7
0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.2 0 0 c0t2d0
0.3 0.1 13.3 1.3 0.0 0.0 0.0 11.2 0 0 c1t3d0
0.3 0.1 13.2 1.3 0.0 0.0 0.0 12.7 0 0 c1t3d0s0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c1t3d0s7
0.3 0.1 13.4 1.3 0.0 0.0 0.0 11.1 0 0 c1t4d0
0.2 0.1 13.3 1.3 0.0 0.0 0.0 12.6 0 0 c1t4d0s0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c1t4d0s7
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 st0

-----------------------------------------------------------------------------

---

References:
iostat -- is there something better?
From: Jay G. Scott
Prev by Date: Re: iostat -- is there something better?
Next by Date: Re: Acrobat reader
Previous by thread: Re: iostat -- is there something better?
Next by thread: Re: iostat -- is there something better?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: iostat -- is there something better?
... Jay G. Scott wrote: ... iostat disklist -l 20 whatever ... so
iostat added 20 additional disks ... (comp.unix.solaris)
Track process
... Below is the iostat information from one of the SUN servers ... (E4500

running
solaris 8 connecting to T3 via SAN switch). ... (comp.unix.solaris)
Re: Why does dd take so long on my Ultra 60
... Angelo Alvarez wrote: ... > starting to fail, so I want to replace
it as soon as possible. ... Does iostat -En show any errors against the drive

you're trying
to ... Scott ... (comp.sys.sun.hardware)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-------
Re: iostat -- is there something better?

-----------------------------------------------------------------------------

---

From: "Daniel Rock" <v200624@xxxxxxxxxxx>
Date: Tue, 13 Jun 2006 18:08:57 +0000 (UTC)

-----------------------------------------------------------------------------

---
Jay G. Scott <gl@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

i've read iostat's man page, maybe that's what i did wrong.
if i do:
iostat disklist -l 20 whatever
it
1. doesn't give me everything in my disklist

Did you use the exact same names as in the "device" column?
"disklist" should also appear *after* any normal options:
iostat -l 20 disklist whatever

2. still cuts the list off at 20

Your disklist probably machted nothing, so iostat added 20 additional disks
to the output.

3, lists disks i don't care about.

See above.

If you use a disklist you should probably set -l 1 (or anything lower than
the number of disks in your disklist).

Example on my machine:

% iostat -xn
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.1 0 0 c0d0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0 0 c0d1
16.7 19.8 1143.1 75.7 0.8 0.1 21.1 3.7 5 8 c3d0
16.0 15.9 1137.6 71.8 0.8 0.2 25.3 4.7 5 10 c4d0
20.3 38.7 1219.0 152.2 0.9 0.2 15.4 3.1 5 12 c2d0
20.3 38.9 1218.6 152.3 0.9 0.2 15.8 3.1 5 12 c1d0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 fd0
6.8 1.1 44.3 5.2 0.0 0.1 0.4 10.6 0 3 md5
0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.7 0 0 md12
0.0 0.0 0.0 0.0 0.0 0.0 0.0 10.1 0 0 md13
0.0 0.0 0.0 0.0 0.0 0.0 0.2 10.9 0 0 md14
3.4 1.1 22.2 5.2 0.0 0.1 0.0 11.4 0 2 md15
3.4 1.1 22.2 5.2 0.0 0.0 0.0 10.6 0 2 md19
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 server:vold(pid367)
% iostat -xn -l 1 c0d1 c1d0 fd0
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0 0 c0d1
20.3 38.9 1218.4 152.3 0.9 0.2 15.8 3.1 5 12 c1d0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 fd0

--
Daniel
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: iostat -- is there something better?
From: Jay G. Scott
References:
iostat -- is there something better?
From: Jay G. Scott
Prev by Date: Re: Acrobat reader
Next by Date: Re: Mount of rootfs(/) is being done in r/w mode
Previous by thread: Re: iostat -- is there something better?
Next by thread: Re: iostat -- is there something better?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: iostat -- is there something better?
... Jay G. Scott wrote: ... iostat disklist -l 20 whatever ... Did
the switches match the names? ... (comp.unix.solaris)
Track process
... Below is the iostat information from one of the SUN servers ... (E4500

running
solaris 8 connecting to T3 via SAN switch). ... (comp.unix.solaris)
SUMMARY: Identifying disks
... After working all night long my mind was not working straight so

obviously I ...
But I had tests fro last night so running iostat with option -n ... or just

take a look
and find disks you are looking for... ... so I do have paths and everything

else but I
need somehow connect them ... (SunManagers)
Re: iostat -- is there something better?
... Daniel Rock wrote: ... iostat disklist -l 20 whatever ... so iostat
added 20 additional disks ... grumble. ... (comp.unix.solaris)
Re: Disk labels (c0t0d0 & c1t0d => c2t0d0 & c3t0d0) weirdness
... Is there a Solaris 9 or 10 command that returns the same ... kind of
info as iostat -En but works for PC disks? ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: iostat -- is there something better?

-----------------------------------------------------------------------------

---

From: gl@xxxxxxxxxxxxxxxxxxxxxxxx (Jay G. Scott)
Date: Thu, 15 Jun 2006 16:04:06 +0000 (UTC)

-----------------------------------------------------------------------------

---
In article <e6muvp$18c6$1@xxxxxxxxxxxxxxx>,
Daniel Rock <v200624@xxxxxxxxxxx> wrote:

Jay G. Scott <gl@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

i've read iostat's man page, maybe that's what i did wrong.
if i do:
iostat disklist -l 20 whatever
it
1. doesn't give me everything in my disklist

Did you use the exact same names as in the "device" column?
"disklist" should also appear *after* any normal options:
iostat -l 20 disklist whatever

2. still cuts the list off at 20

Your disklist probably machted nothing, so iostat added 20 additional disks
to the output.

3, lists disks i don't care about.

See above.

If you use a disklist you should probably set -l 1 (or anything lower than
the number of disks in your disklist).

aaugh! that helps.

yours and the other posters comments helped, but there's still a problem:

lurch# iostat -mnxszM -l 1 `cat /tmp/rot3` 20 2
extended device statistics
r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0 0 d19
0.0 0.3 0.0 0.0 0.0 0.0 7.7 11.5 0 0 d37 (/var)
0.0 0.6 0.0 0.0 0.0 0.0 9.6 28.2 1 1 d38 (/opt)
0.6 7.7 0.0 0.1 0.0 0.0 0.8 0.9 1 1 d44
0.0 1.2 0.0 0.0 0.0 0.0 4.8 0.7 1 0 d63

d19, for example, doesn't show the mount point. grumble.

but thanks to all.

(ahhh.... my complaining about where the disk label was, left, right,
is because if i were to write some script to digest the output it would
be harder to change options.)

[snip]

--
Daniel

--
Jay Scott 512-835-3553 gl@xxxxxxxxxxxxxxxx
Head of Sun Support, Sr. Operating Systems Specialist
Applied Research Labs, Computer Science Div. S224
University of Texas at Austin
.

-----------------------------------------------------------------------------

---

References:
iostat -- is there something better?
From: Jay G. Scott
Re: iostat -- is there something better?
From: Daniel Rock
Prev by Date: Re: Solaris 8 system duplicate IP address
Next by Date: How to compile Apache 2.2.2 on Solaris 10 ?
Previous by thread: Re: iostat -- is there something better?
Next by thread: Memory for Ultra 45
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
SUMMARY: Identifying disks
... After working all night long my mind was not working straight so

obviously I ...
But I had tests fro last night so running iostat with option -n ... or just

take a look
and find disks you are looking for... ... so I do have paths and everything

else but I
need somehow connect them ... (SunManagers)
Re: Disk labels (c0t0d0 & c1t0d => c2t0d0 & c3t0d0) weirdness
... Is there a Solaris 9 or 10 command that returns the same ... kind of
info as iostat -En but works for PC disks? ... (comp.unix.solaris)
RE: [Q] how to check disk performance ?
... I'm sorry maybe I'm missing something here, ... >> We would like check
disks performance under AS2.1. ... > iostat is a pretty simple disk utility

which
is for some reason missing ... (RedHat)
RE: [Q] how to check disk performance ?
... > We would like check disks performance under AS2.1. ... > disks and SAN
disks. ... Sysstat - The sar, mpstat, and iostat commands ... (RedHat)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

2、Re: Mount of rootfs(/) is being done in r/w mode

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Tue, 13 Jun 2006 16:52:01 GMT

-----------------------------------------------------------------------------

---
Imran <imran.1.618@xxxxxxxxx> wrote:

Hi all,

I was trying to write a service that checks attributes of rootfs before
it is
checked by fsck. The service has a dependency on boot-archive and has a
dependent "/system/filesystem/usr" service. It seems "/" is being
mounted in r/w mode before a
check is done on it. is that correct?

No, not normally.

Realize that the root filesystem initial (read-only) mount is
constructed directly by the kernel. There's no startup script that does
so, because it has to already be mounted to read the script.

If you're interested in '/', why do you have a dependency on usr?

When is the rootfs first mounted in the boot process?

A mount is constructed directly by the kernel as it boots.

It is mounted read/write by svc:/system/filesystem/root:default.

Is it ever
mounted in
read-only mode

yes. The initial kernel mount is read only. A pre-solaris 10 'boot -b'
will give you a shell before root is mounted read-write.

For Solaris 10, use 'boot -m milestone=none'.

I don't know what 'mount' is looking at. It will report / as
read/write, but in fact it is still read-only at that point.

or is Solaris fsck capable of checking filesystems that
are
mounted w/o damaging them?

no.

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Mount of rootfs(/) is being done in r/w mode
From: Andrew Gabriel
References:
Mount of rootfs(/) is being done in r/w mode
From: Imran
Prev by Date: Re: Memory for Ultra 45
Next by Date: Re: iostat -- is there something better?
Previous by thread: Mount of rootfs(/) is being done in r/w mode
Next by thread: Re: Mount of rootfs(/) is being done in r/w mode
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: usb-storage && iRIVER flash player problem
... > The device is seen, but there is no way to mount it, nor re-partition

...
explain that the mountable vfat partition under linux is the ... > debugging

enabled
and not as modules, directly into the kernel. ... > nfs warning: ...

(Linux-Kernel)
Re: Getting CDRW to work in Debian
... Note that these instructions match a tuned modular kernel such as ...

method
to install your network and sound drivers. ... Test your configuration using

cdrecord
-scanbus. ... Create mount points for these devices. ... (Debian-User)
Re: XFS Mount Hangs the Partition (on latest kernel + many old 2.6.x ones)
... > attempting to mount a certain XFS partition (but not a different one on

the ...
> same hard-disk) caused the mount process to hang, ... > (that was using

kernel
2.4.15-rc2 from Linus), and then rebooted twice, ... (Linux-Kernel)
PROBLEM: (Bug report) USB Mass Storage in 2.6.0-test11 (cannot mount flash

drive)
... I cannot mount any USB mass storage drive when using kernel 2.6.0-test11

...
Dec 6 19:28:18 mycomp modprobe: FATAL: Module usb_storage already in kernel.

... Dec 6
14:49:25 mycomp kernel: Initializing USB Mass Storage driver... ...

4000-40ff: PCI CardBus
#02 ... (Linux-Kernel)
Re: evangelizing linux not needed
... Both should be umask (do a "man mount" to see mounting options). ...
an installer to guess things for you. ... used for your stock kernel. ...
(alt.os.linux)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: Mount of rootfs(/) is being done in r/w mode

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Wed, 14 Jun 2006 23:50:34 GMT

-----------------------------------------------------------------------------

---
Andrew Gabriel <andrew@xxxxxxxxxxxxxxxxxxxx> wrote:

For Solaris 10, use 'boot -m milestone=none'.

I don't know what 'mount' is looking at. It will report / as
read/write, but in fact it is still read-only at that point.

Maybe you have an old /etc/mnttab file underneath the mnttab
mount point, and mnttab wasn't yet mounted?

I thought of that right after I got the machine back up, but didn't want
to set everything back up again. So I finally got around to trying
another test and that doesn't seem to be the case...

# umount /etc/mnttab
# df -k
# ls -l /etc/mnttab
-r--r--r-- 1 root root 0 Jan 21 2005 /etc/mnttab

So it really is the kernel view of the mounts, not a stale file.

The "early" mnttab has no flag indications at all (neither rw or ro).

# head -1 /etc/mnttab
/pci@1f,0/pci@1,1/ide@3/disk@0,0:a / ufs dev=2200008 0

So I think 'mount' just assumes that no flags equals read/write (which
isn't true in this case).

In fact.... I don't know how to mount this filesystem (explicitly) read
only. Now that I look, the mount_ufs man page doesn't show a 'ro'
option, and my attempts to do so fail.

# mount -r -F ufs -o remount /dev/dsk/c0t0d0s0 /
mount: /dev/dsk/c0t0d0s0 is not this fstype
# mount -F ufs -o remount,ro /dev/dsk/c0t0d0s0 /
mount: /dev/dsk/c0t0d0s0 is not this fstype
but...
# mount -F ufs -o remount /dev/dsk/c0t0d0s0 /
#

"normal" filesystems can be mounted read-only, so why were the ro | rw
options removed from the mount_ufs man page?

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

References:
Mount of rootfs(/) is being done in r/w mode
From: Imran
Re: Mount of rootfs(/) is being done in r/w mode
From: Darren Dunham
Re: Mount of rootfs(/) is being done in r/w mode
From: Andrew Gabriel
Prev by Date: T2000 running Clearcase
Next by Date: Re: Acrobat reader
Previous by thread: Re: Mount of rootfs(/) is being done in r/w mode
Next by thread: Customization Of Java Desktop
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: HFS Plus R/W Mounting Problem
... > I'm trying to mount my iPod on FC2 as a read/write hfs plus file ... My
fstab entry looks like: ... > Sometimes it mounts as read/write but other

times
its just read only. ... What kernel are you using? ... (Fedora)
Re: mount & umount read-only floppy: unmount failed: Input/output error ??
... Rob wrote: ... if the mount didn't work. ... CAN read/write
mount a read/only-floppy. ... Unless, you take the floppy out, change it ...
(freebsd-questions)
HFS Plus R/W Mounting Problem
... I'm trying to mount my iPod on FC2 as a read/write hfs plus file system.

...
fstab entry looks like: ... (Fedora)
RE: Freebsd ufs under WIN
... > for read/write. ... > I am almost 100% certain that the real question
was: ... >> not only can I read and write to them, I can mount them as ...
is that I can *map* Samba shares to Network drives, ... (freebsd-questions)
Re: why vnode as readonly fs?
... > sam wrote: ... > You can't mount a cd9660 fs read/write.
... > matter of mdconfig but of the fs itself. ...

(comp.unix.bsd.freebsd.misc)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

3、Re: 3310 RAID1 (mirror) Question

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 13 Jun 2006 08:49:50 GMT

-----------------------------------------------------------------------------

---
In article <1150166605.509903.285610@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"BC" <unix_ed@xxxxxxxxx> writes:

Question about setting up a Storedge 3310 - mirror. The 3310 has 12
drives, 2 channels. When creating the logical drive, I select 5 drives
from each channel (10 drives total) and select RAID1 (2 drives will be
saved for Global spares). I understand that the array will configure
RAID 1+0 and is also smart enough to setup the mirrored pairs across
both channels as described in this short Sun Doc:
http://sunsolve.sun.com/search/document.do?assetkey=1-9-71182-1

3 hrs later, after the drives are initialized, I see one logical drive,
but instructions are not clear how to setup the Solaris 9 OS to use
both controllers? I'm not sure how the dual channels are presented to
the OS? Or does the array deal with it all?

The 3310 will only give you a single host connection. (You could
connect another host to the array, or even the same host more than
once, but there isn't any way to get path redundancy and failover.)

Given two controllers, if one fails then it will just use the
other. Automatically and transparently. Works great. No interruption to
service.

(The downside to this is that you have to actively monitor the array to
check that both controllers are running, because the system might not
notice at all if one fails.)

Should I have setup 5 drives (channel 1) as RAID 0 and the other 5
drives as RAID 0 (channel 3), presented them both to the server and
used SVM to mirror across the controllers?

Definitely not. If you lose a disk in a stripe you lose the whole
stripe and all your data redundancy goes. If you want 0+1 to turn into
1+0, you need to do it all inside the array or all inside SVM.

Server is a V480 (Solaris 9), 3310 array firmware version 3.25, Fiber
channel

That doesn't make sense. The 3310 is a SCSI array, if you had Fiber
Channel it would be a 3510 (given the channels IDs you mention, it
sounds like you do have a SCSI based 3310). But if you did have a 3510
then my notes are here:

http://www.petertribble.co.uk/Solaris/3510.html
http://www.petertribble.co.uk/Solaris/mpxio.html

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: 3310 RAID1 (mirror) Question
From: BC
References:
3310 RAID1 (mirror) Question
From: BC
Prev by Date: Re: Problem with SSH
Next by Date: Re: Acrobat reader
Previous by thread: 3310 RAID1 (mirror) Question
Next by thread: Re: 3310 RAID1 (mirror) Question
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: 3310 RAID1 (mirror) Question
... connect another host to the array, or even the same host more than ... I

select
5 drives ... I understand that the array will configure ... both channels
as described in this short Sun Doc: ... (comp.unix.solaris)
Re: Single P4, many IDE PCI cards == trouble??
... * Promise 20268 PCI cards ... Only 3 drives, ... and one disk
to each of the channels of the SI680 ... (Linux-Kernel)
Re: New Hd Install - No Udma Support, Dma 2 Only
... and secondary channels in device manager as well as the ide controller.

...
No listing of UDMA type. ... Does your motherboard really support large hard

drives?
... (alt.comp.hardware.pc-homebuilt)
Re: Booting
... my problem is that I have three hard drives all set ... All on their own

channels
and so forth set to master. ... > the problem because they are all master

drives, ...
(microsoft.public.windowsxp.general)
3310 RAID1 (mirror) Question
... I select 5 drives ... RAID 1+0 and is also smart enough to setup the

mirrored
pairs across ... both channels as described in this short Sun Doc: ... used
SVM to mirror across the controllers? ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: 3310 RAID1 (mirror) Question

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 13 Jun 2006 13:58:37 GMT

-----------------------------------------------------------------------------

---
In article <1150204064.354224.221870@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"BC" <unix_ed@xxxxxxxxx> writes:

Thanks for the help Peter,
You're right, the 3310 is SCSI and not fiber. (You know how it goes
whenit's late).

Yup. ;-)

you stated:

The 3310 will only give you a single host connection. (You could
connect another host to the array, or even the same host more than
once, but there isn't any way to get path redundancy and failover.)

Here's what I see with a format though:

0. c1t0d0 sec 424>
/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w2100000c50ebb5f7,0
1. c1t1d0
/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w21000011c62a46e2,0
2. c4t1d0
/pci@8,700000/pci@3/scsi@5/sd@1,0
3. c6t1d0
/pci@8,700000/pci@2/scsi@5/sd@1,0

So you have two SCSI connections to the array?

4. c8t40d0
/pci@8,700000/SUNW,qlc@4/fp@0,0/ssd@w216000c0ff80670d,0

Items 2 and 3. I have two LUN's - 682GB each. I assume one LUN from
each side of the 3310. Based on what you stated, I would expect to
see one LUN. Do I now use SVM and create submirrors between 2 and 3?

How many logical drives have you created? You say you've created just
the one (mirror with 10 drives - which would be the right size if you
have 146G drives). What does the sccli "show ld" command give you?

If you've just created the one logical drive then you're just seeing
the same LUN twice - by different paths. In that case, really bad
things can happen if you try and access it from both paths. (As far as
I know, it isn't possible to use mpxio to use both paths in the same
way you can for a fiber array.)

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: 3310 RAID1 (mirror) Question
From: BC
References:
3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Prev by Date: Re: Sun Google
Next by Date: Re: Sun Google
Previous by thread: Re: 3310 RAID1 (mirror) Question
Next by thread: Re: 3310 RAID1 (mirror) Question
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: IBM FastT vs. EMC Clarion
... AIX hosts have to be shutdown when doing the firmware upgrade ... any AIX
servers attached to that unit will have to be shutdown. ... interrupting host

I/O access to
the array. ... interrupting host I/O access to the array thousands of times

with
many ... (AIX-L)
Re: 3310 RAID1 (mirror) Question
... connect another host to the array, or even the same host more than ... I

select
5 drives ... I understand that the array will configure ... both channels
as described in this short Sun Doc: ... (comp.unix.solaris)
Re: Problem with filesystem on RAID5 array (fsck -b?)
... Dell PowerVault 220S disk cabinet. ... Two days ago a disk failed in
one of the arrays. ... Rebuilding the array ... delete logical drives

supported
by controller ... (freebsd-stable)
Re: Newbie: graphics cards
... I am using an IBM 3H/3L SCSI Raid Controller Card and the system sees

that, but does
not see my RAID Array. ... have you booted the ServeRAID support CD to set up

...
your array and logical drives? ... (comp.sys.ibm.pc.hardware.misc)
Re: Mirror two A1000 with two (non-cluster) nodes
... Each array has two RAID-5 volumes, for a total of 4 volumes: ...

Controller
A owns host 1's ... while Controller B owns host 2's volume. ... You can
use pigtail adapters, where the connector is on one side of the ...

(comp.sys.sun.admin)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-------
Re: 3310 RAID1 (mirror) Question

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 13 Jun 2006 22:03:18 GMT

-----------------------------------------------------------------------------

---
In article <1150228835.650805.248580@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"BC" <unix_ed@xxxxxxxxx> writes:

RE: sccli
I downloaded and installed sccli, but can't get info on the 3310
because it's not the primary controller. Most of the sccli commands
require the primary controller. As I showed in previous output a few
posts back, the 3310 configured itself on the secondary controller. In
the "Edit Logical Drives" section, there is a submenu named "Logical
Drive Assignments". It has an option to:
"Redundant Controller Logical Drive Assign to Primary Controller?" Yes
or No

when I select Yes, it changes my host Luns so they are on the primary
and not the secondary chl- but requires a controller reset. I was a
bit afraid it was going to change my disk logical device names in the
OS, so I bailed out of it and set things back to the secondary
controller.

My question is: If I change to Primary (so I can use sccli sub
commands), will my logical device name for the drive (as it appears in
the format utility) change? Thus, destroying my exisiting file systems
and soft partitions?

Now, I don't know what's going to happen. I've never tried what you're
thinking of doing, so I'm not going to do anything more than take a
guess based on my understanding of how the thing works.

My guess is that if you move it from the secondary to the primary
controller then it will change the ID from t1 to t0 (so c4t1d0 becomes
c4t0d0). It won't destroy any filesystems or anything like that, but
you would have to go through the system so that Solaris knows to look
in a different place.

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: 3310 RAID1 (mirror) Question
From: Joe D.
References:
3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: BC
Prev by Date: Re: Customization Of Java Desktop
Next by Date: Re: changing routing table preference
Previous by thread: Re: 3310 RAID1 (mirror) Question
Next by thread: Re: 3310 RAID1 (mirror) Question
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: 3310 RAID1 (mirror) Question
... I downloaded and installed sccli, but can't get info on the 3310 ...

because
it's not the primary controller. ... On my 3510s with two logical drives, ...
(comp.unix.solaris)
MVP? how do I restore video during boot
... the system wanted to treat the primary controller as the secondary

controller
... So when we boot - all goes well from the bios boot through windows boot

...
(microsoft.public.windowsxp.perform_maintain)
Re: 3310 RAID1 (mirror) Question
... Just install the sccli software on the attached host. ... is on the

primary
controller. ... On my 3510s with two logical drives, ... assigned one to each
controller. ... (comp.unix.solaris)
Configure multipathing in 3510 with V440
... First logical disk I assigned to primary controller. ... But when I

remove
primary FCAL channel cable from primary controller. ... (SunManagers)
Re: Sandisk fixed disk utility error #20
... (for a disk on the primary controller) ... atcfwchg /P /F ...
(for a disk on the secondary controller) ...

(microsoft.public.windowsxp.embedded)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
------
Re: 3310 RAID1 (mirror) Question

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 13 Jun 2006 22:03:18 GMT

-----------------------------------------------------------------------------

---
In article <1150228835.650805.248580@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"BC" <unix_ed@xxxxxxxxx> writes:

Now, I don't know what's going to happen. I've never tried what you're
thinking of doing, so I'm not going to do anything more than take a
guess based on my understanding of how the thing works.

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

-----------------------------------------------------------------------------

---

Relevant Pages
Re: 3310 RAID1 (mirror) Question
... I downloaded and installed sccli, but can't get info on the 3310 ...

controller
... So when we boot - all goes well from the bios boot through windows boot

...
(microsoft.public.windowsxp.perform_maintain)
Re: 3310 RAID1 (mirror) Question
... Just install the sccli software on the attached host. ... is on the

(microsoft.public.windowsxp.embedded)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: 3310 RAID1 (mirror) Question

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 14 Jun 2006 16:13:59 GMT

-----------------------------------------------------------------------------

---
In article <1150288691.925660.262540@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"Joe D." <newbie_from_newbie@xxxxxxxxx> writes:

Good morning, gentlemen! A very interesting post.

I am about to set up a 3310 (probably a 3320, actually, but I presume
they are essentially the same as far as config goes?).

Pretty well. The 3320 is Ultra320, while the 3310 is Ultra160. And they
do take different firmware (new patches for both were released today)
so don't try applying 3310 firmware to a 3320 (or vice versa). But as
far as configuration goes I haven't noticed any differences.

I presume from
the post that you both have the RAID controller option. We're not going
to go with the RAID controller, due to cost considerations. Rather, I
was planning on simply using SVM for the system's boot disks, and use
the 3310's disks via raw access for a Sybase database.

I wonder if I could trouble you to elaborate a little on some points
made earlier:

Peter Said:
The 3310 will only give you a single host connection. (You could
connect another host to the array, or even the same host more than
once, but there isn't any way to get path redundancy and failover.)

Given two controllers, if one fails then it will just use the
other. Automatically and transparently. Works great. No interruption to
service.

and later when BD noted that he had 2 SCSI connections to the array and
was seeing 2 disks:

.... disconnect the SCSI cable from the
host. You don't really want this to be accessible, as it's just asking
for trouble.

One of the things I was hoping for in purchasing these arrays was the
dual-path redundancy (which it sounds like I'm not gonna get).

Well, SCSI is intrinsically single-path - unlike FC disks which are
dual-ported. So with JBOD you could have one path to a disk and a
different path to its mirrored (on the host) partner.

Is this
to say then, that even if I order a pair of PCI SCSI cards that I
shouldn't connect both of them?

With a RAID controller, you shouldn't. With JBOD, you should, and split
the busses so you attach half the disks to one card and half to the other.

My past experience had been with
Veritas which recognized automatically 2 paths to the same disk(s) and
presented you with only the one, even though you could see both from
format.

That's a good point. I hadn't thought about Veritas (previous
experience makes my mind blank it out - oh the pain!). I don't actually
know what Veritas would do with it.

I realize we may be talking apples and oranges here since you're using
the RAID controller (again, I'm presuming) and we won't be, and also
we'll be accessing them as raw devices, but is what I'm planning
feasible/recommended? Any alarm bells going off in your heads?

If you take out the raid controller, then I would run with a dual-bus
configuration. Connect one half of the disks to one card, the other
half to the other card, and then create mirrors (either SVM or within
Sybase) between disks on the two controllers.

In other words, split it in half and treat it as two 6-disk arrays that
happen to be in one box.

Note that the rules for JBOD and the rules for RAID are different.

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: 3310 RAID1 (mirror) Question
From: Joe D.
References:
3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: BC
Re: 3310 RAID1 (mirror) Question
From: Peter Tribble
Re: 3310 RAID1 (mirror) Question
From: Joe D.
Prev by Date: Re: Faster backups - possibly Solaris pipe buffer size
Next by Date: Power Management Problems (DPMS)
Previous by thread: Re: 3310 RAID1 (mirror) Question
Next by thread: Re: 3310 RAID1 (mirror) Question
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
MPXIO: Cant get to encapsulate drives ...
... I can't get MPXIO to encapsulate the following pairs of disks ... are

derived
from a Clariion FC4700 unit (through two Brocade ... # Load balancing global

configuration:
... # List of ULP modules for loading during port driver attach time ...

(comp.unix.solaris)
Re: Flashing lights on my disk drives on an Alphaserver ES40
... When you have internal disks a junk card is installed in the ES40 -

assuming ...
> or whatever the controller is - even if it is not a RAID controller it>

should be listed
in SRM e.g. a PKxn device. ... > If the RAID controller is seen by SRM, go

into the RAID config
utility> and check that your physical drives are seen and your> logical

drives are still
defined ... > AND a RAID card, it might be normal to have no devices showing

on the>
internal SCSI bus, so if the RAID card disappears then there would> be

'unused' controllers - make
sure you follow the disk cables back! ... (comp.os.vms)
Re: Flashing lights on my disk drives on an Alphaserver ES40
... When you have internal disks a junk card is installed in the ES40 -

assuming ...
> or whatever the controller is - even if it is not a RAID controller it>

should be listed
in SRM e.g. a PKxn device. ... > If the RAID controller is seen by SRM, go

into the RAID config
utility> and check that your physical drives are seen and your> logical

drives are still
defined ... > AND a RAID card, it might be normal to have no devices showing

on the>
internal SCSI bus, so if the RAID card disappears then there would> be

'unused' controllers - make
sure you follow the disk cables back! ... (comp.sys.dec)
Re: Odd news transit performance problem
... > box is idle - there's no significant contention for memory, disk, ...

The
host is fed news from only two sources which on average equates to 25 ...

This still only
pushes the disks to 50% busy. ... 11424 Kbytes allocated to network ...

(freebsd-performance)
Re: ext3 filesystem not working with raid5
... and the array might already have reconstructed with bad data due ...

consider
*maxtor*250GIG IDE disks to be nice for single-user ... -rf' or a rooted

system or a RAID
controller running amuck at startup. ... > exactly what the problem is with

the ext3
filesystem and how to go ... (alt.linux)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

4、Re: Optimizing memory usage of high performance webserver

-----------------------------------------------------------------------------

---

From: Rich Teer <rich.teer@xxxxxxxxxxxxxx>
Date: Sun, 11 Jun 2006 18:28:36 GMT

-----------------------------------------------------------------------------

---
On Sun, 11 Jun 2006, fredje110@xxxxxxxxxxx wrote:

We are running 4 Sun Fire V440 with 4 CPU's and 8 GB RAM as Apache
webservers. We do not use a dedicated caching/SQUID tier and we do not
have dedicated image servers (the amount of pictures on our website is
very large). It should not be any surprise that our internal discs have
reached their I/O limits.

iostat will give you some more info here.

On the other site the output of top shows the following:

Memory: 8192M real, 4892M free, 177M swap in use, 9538M swap free

It would make a big difference if we could load 3-4 GB of cached
content into memory (RamDisk?)

Top isn't the best tool to use on Solaris (use prstat instead), but
more to the point, if your files are static, they'll be cached in
RAM by the kernel anyway (more pressing needs for RAM nothwithstanding).

Our web tier is running Sun Solaris 8. How could we tune/convince
Solaris to load more cached content into memory and reduce disk I/O.

Use performance analysis tools like iostat to make sure you are actually
disk I/O bound before you do any tuning. But I will say this: you will
very likely see a significant performance increase if you upgrade those
servers to Solaris 10. Upgrading to Solaris 10 will also enable you to
use DTrace to help identify any performance bottlenecks. (If you can
wait a week or two, you'll be able to use SOlaris 10 6/06, which comes
with ZFS...)

HTH,

--
Rich Teer, SCNA, SCSA, OpenSolaris CAB member

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-group.com/rich
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Optimizing memory usage of high performance webserver
From: KJ
Re: Optimizing memory usage of high performance webserver
From: Chris Lawson
Re: Optimizing memory usage of high performance webserver
From: Daniel Rock
References:
Optimizing memory usage of high performance webserver
From: fredje110
Prev by Date: Re: New server to play on
Next by Date: Re: Optimizing memory usage of high performance webserver
Previous by thread: Optimizing memory usage of high performance webserver
Next by thread: Re: Optimizing memory usage of high performance webserver
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(09)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: Optimizing memory usage of high performance webserver

-----------------------------------------------------------------------------

---

From: "victor.engle@xxxxxxxxx" <victor.engle@xxxxxxxxx>
Date: 11 Jun 2006 13:30:59 -0700

-----------------------------------------------------------------------------

---
Daniel,

I agree with Rich's comments. We need to see the iostat data for those
drives housing your content. Also some network and cpu statistics and
maybe some vmstat data captured during a time when the systems is
experiencing the performance issue.

The V440 I believe has the Sparc IIIi processors so depending on how
much network bandwidth you're trying to support the system could even
be CPU bound at times. Is the content storage direct attached? What
kind of storage is it. A fibre channel SAN of course would be very
expensive but if you have to have faster storage then a midrange SAN
solution might be a reasonable thing to consider but I wouldn't insist
on active/active. That would likely add significant cost and it is
really hard to believe it's needed for a web server. You can configure
an active passive array such that you can utilize 2 or more HBA's on
the host side. I can provide details if needed.

Also, a selling point for the SAN idea is that it would open up shared
storage clustering possibilities for you.

Good luck with resolving the problem.

Regards,

Vic Engle

Daniel Rock wrote:

Rich Teer <rich.teer@xxxxxxxxxxxxxx> wrote:

On Sun, 11 Jun 2006, fredje110@xxxxxxxxxxx wrote:

Memory: 8192M real, 4892M free, 177M swap in use, 9538M swap free

Top isn't the best tool to use on Solaris (use prstat instead),

prstat won't display anything similar to that above (physical memory usage),
so why use it if it cannot give me the information I'm interested in?

--
Daniel

-----------------------------------------------------------------------------

---

References:
Optimizing memory usage of high performance webserver
From: fredje110
Re: Optimizing memory usage of high performance webserver
From: Rich Teer
Re: Optimizing memory usage of high performance webserver
From: Daniel Rock
Prev by Date: Re: Optimizing memory usage of high performance webserver
Next by Date: Re: Optimizing memory usage of high performance webserver
Previous by thread: Re: Optimizing memory usage of high performance webserver
Next by thread: Re: Optimizing memory usage of high performance webserver
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Information store on WSS 2003
... communication with the storage, not with the network. ... Check for

firmware
and driver updates for storage solution. ... Check with your storage

manufacturer to see
if one or more settings ... Please do not send email directly to this alias.

...
(microsoft.public.exchange.admin)
Re: [SLE] Why logical volume?
... I thought we were talking storage not replacing NIC cards. ... cents
at the local computer store here, so yes a retool of a NETWORK ... >

site-wide spam
filters at catherders.com. ... (SuSE)
Re: Shared storage question.
... "A SAN is also referred to as a Storage Area Network. ...

network-attached
storage server is attached to another server through the ...

(microsoft.public.windows.server.clustering)
[PAPER] Juggling with packets: floating data storage
... properties of the Internet or any other large network to create ...

volatile
distributed data storage of a large capacity. ... As a typical SMTP server

prevents
to relay mail from the client to ... (Bugtraq)
[VulnWatch] [PAPER] Juggling with packets: floating data storage
... properties of the Internet or any other large network to create ...

volatile
distributed data storage of a large capacity. ... As a typical SMTP server

prevents
to relay mail from the client to ... (VulnWatch)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(09)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
---
Re: Optimizing memory usage of high performance webserver

-----------------------------------------------------------------------------

---

From: ptribble@xxxxxxxxxxx (Peter Tribble)
Date: 11 Jun 2006 18:32:31 GMT

-----------------------------------------------------------------------------

---
In article <1150046519.938856.195290@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
fredje110@xxxxxxxxxxx writes:

Hello!

Actually, this isn't obvious. Disks are fairly fast, networks
(affordable wide-area ones anyway) somewhat less so, high-traffic web
sites stress the TCP/IP stack.

On the other site the output of top shows the following:

Memory: 8192M real, 4892M free, 177M swap in use, 9538M swap free

It would make a big difference if we could load 3-4 GB of cached
content into memory (RamDisk?)

Our web tier is running Sun Solaris 8. How could we tune/convince
Solaris to load more cached content into memory and reduce disk I/O.

It does that anyway. Solaris will use all free memory as a disk
cache. So you already have a disk cache of 5G or so (based on the top
output).

Assuming that you have some basis for your 3-4G number, then that
should be the effective working set and everything should be cached in
memory already.

What do your apache traffic, network traffic, and disk I/O actually
look like? Request rate, average request size, network utilization,
disk read/writes and transfer sizes? Do you know whether it's file I/O
or directory lookups that are loading your disks?

You could try Solaris NCA, which is a direct kernel cache.

--
-Peter Tribble
L.I.S., University of Hertfordshire - http://www.herts.ac.uk/
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
.

-----------------------------------------------------------------------------

---

References:
Optimizing memory usage of high performance webserver
From: fredje110
Prev by Date: Re: Optimizing memory usage of high performance webserver
Next by Date: Acrobat reader
Previous by thread: Re: Optimizing memory usage of high performance webserver
Next by thread: Re: Optimizing memory usage of high performance webserver
Index(es):
Date
Thread

随机文章：

0511月1，2，3日 2005-12-19

0509月28，29，30日 2005-11-09

0509月1，2，3日 2005-10-08

0507月13，14，15日 2005-08-30

0506月16，17，18日 2005-08-06

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0606月1，2，3，4，5，6日

mmmmn — Sun, 30 Jul 2006 01:38:58 +0800

1、Re: File names change to upper case in USB drive

-----------------------------------------------------------------------------

---

From: Robert Lawhead <news0000.5.unixguru@xxxxxxxxxxxxxxx>
Date: Wed, 07 Jun 2006 21:14:09 -0700

-----------------------------------------------------------------------------

---
Keith Thompson wrote:

Robert Lawhead <news0000.5.unixguru@xxxxxxxxxxxxxxx> writes:

googler wrote:

This problem is strange. I am using a San Disk Mini Cruzer USB drive
(512 MB) to copy some files from a Windows system to a Solaris 10
system. After copying the files in the USB drive from the Windows
system, when I plug it into the Solaris system, I see that many (but
not all) file and directory names that were actually in lowercase have
changed to uppercase. So the file aaa becomes AAA.

Any idea why this is happening? How to fix this?

Thanks.

"man pcfs". Perhaps you should mount using PCFS_MNT_FOLDCASE option.

That causes all file names to be mapped to lower case, which probably
isn't what's wanted.

Normally FAT filesystems are case-preserving; files retain the case
with which they were created, though you can't have two files in the
same directory whose names differ only in case, and you can refer to a
file "foo.txt", as "FOO.TXT, etc. Perhaps there's a bug in the
Solaris implementation of pcfs?

A good workaround is to copy a zip file or tarball to the USB drive,
rather than copying the files directly. That way the contained file
names are stored in the file itself rather than in the filesystem.

Ken,
The following excerpt from the solaris express pcfs (7) man
page seems (at least to me) to account for the observation made by
the OP. I agree PCFS_MNT_FOLDCASE may not do just what OP wants...
Had he asked how to preserve names in transfer, rather than how to
supress peculiar case wrap of file names already on the USB drive,
I would have offered the same advise as you.

Bob
File Systems pcfs(7FS)

When creating a file name, pcfs creates a short file name if
it fits the DOS short file name format, otherwise it
creates a long file name. This is because long file names
take more directory space. Because the root directory of a
pcfs file system is fixed size, long file names in the root
directory should be avoided if possible.

When displaying file names, pcfs shows them exactly as they
are on the media. This means that short names are displayed
as uppercase and long file names retain their case. Earlier
versions of pcfs folded all names to lowercase, which can be
forced with the PCFS_MNT_FOLDCASE mount option. All file
name searches within pcfs, however, are treated as if they
were uppercase, so readme.txt and ReAdMe.TxT refer to the
same file.
.

-----------------------------------------------------------------------------

---

References:
File names change to upper case in USB drive
From: googler
Re: File names change to upper case in USB drive
From: Robert Lawhead
Re: File names change to upper case in USB drive
From: Keith Thompson
Prev by Date: Re: File names change to upper case in USB drive
Next by Date: Re: max. size limit for mblk's
Previous by thread: Re: File names change to upper case in USB drive
Next by thread: Re: File names change to upper case in USB drive
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: PEZ MP3 Player vs. Solaris 9(SPARC)
... use with my Sun Blade 100 running Solaris 9 4/04 with the latest ...

enable
debug output for the pcfs filesystem, ... Trying to mount it as a pcfs

filesystem
... pcfs for fat media with 2048 bytes/sector. ... (comp.unix.solaris)
Re: File names change to upper case in USB drive
... to copy some files from a Windows system to a Solaris 10 ... After

copying
the files in the USB drive from the Windows ... system, when I plug it into

the Solaris system,
I see that many (but ... Normally FAT filesystems are case-preserving; ...

(comp.unix.solaris)
Re: File names change to upper case in USB drive
... googler wrote: ... to copy some files from a Windows system to a Solaris
10 ... After copying the files in the USB drive from the Windows ... system,

when
I plug it into the Solaris system, I see that many (but ...

(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-----
Re: File names change to upper case in USB drive

-----------------------------------------------------------------------------

---

From: js@xxxxxxxxxxxxxxx (Joerg Schilling)
Date: 8 Jun 2006 20:59:47 GMT

-----------------------------------------------------------------------------

---
In article <lnejy0ibjn.fsf@xxxxxxxxxxxxxxx>,
Keith Thompson <kst-u@xxxxxxx> wrote:

Robert Lawhead <news0000.5.unixguru@xxxxxxxxxxxxxxx> writes:

googler wrote:

Any idea why this is happening? How to fix this?

This is a Microsoft bug.

M$ ignores the case and even lists filenames that fit into 8.3 as
lowercase.

M$ uncorrectly creates FAT uppercase names for a lowercase filename
that fits into 8.3

If you use Solaris to create the filenames, Solaris correctly uses VFAT
names ;-)

--
EMail:joerg@xxxxxxxxxxxxxxxxxxxxxxxxxxx (home) Jörg Schilling D-13353 Berlin
js@xxxxxxxxxxxxxxx (uni)
schilling@xxxxxxxxxxxxxxxxxxx (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
.

-----------------------------------------------------------------------------

---

References:
File names change to upper case in USB drive
From: googler
Re: File names change to upper case in USB drive
From: Robert Lawhead
Re: File names change to upper case in USB drive
From: Keith Thompson
Prev by Date: Re: Sun changes patch policy
Next by Date: Solaris 9
Previous by thread: Re: File names change to upper case in USB drive
Next by thread: Re: Sol 10 x86 not recognizing NIC
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: A word of caution
... I post from the Windows system because I can do it remotely. ... No use

trying
to synchronize what I've read/posted on the Mac system with that of the

Windows system. ...
Use Remote Desktop ... I'm actually thinking of loading x86 Solaris onto my

laptop.
... (comp.sys.mac.advocacy)
File names change to upper case in USB drive
... I am using a San Disk Mini Cruzer USB drive ... to copy some files from a

Windows
system to a Solaris 10 ... system, when I plug it into the Solaris system, I

see that
many (but ... (comp.unix.solaris)
Re: File names change to upper case in USB drive
... to copy some files from a Windows system to a Solaris 10 ... After

copying
the files in the USB drive from the Windows ... system, when I plug it into

the Solaris system,
I see that many (but ... Normally FAT filesystems are case-preserving; ...

when
I plug it into the Solaris system, I see that many (but ...

(comp.unix.solaris)
Possible to burn CD that is readable on both Windows & Solaris?
... I would like to burn some data CDs that will be readable on both Windows

...
The files have long filenames. ... including the long filenames, but it won't

mount on
Solaris, either ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

2、Re: High TCP connect timeout rate questions

-----------------------------------------------------------------------------

---

From: Sam N <sun@xxxxxxxxxxxxxx>
Date: Thu, 08 Jun 2006 07:23:09 GMT

-----------------------------------------------------------------------------

---
Kyle Tucker wrote:
I have received this error on a port listened to by a Squid server. Is this

warning issued prior to reaching the point
where pending TCP requests would have been dropped? And if
that point is reached, would the counter for tcpListenDropQ0 shown in netstat

be greater than 0? Mine is at 0. Thanks.

Hi Kyle

It sounds like maybe the box is getting hammered - Is it really busy? There

are steps you can take to tune the stack on Solaris boxes, you can make

changes to set numerous variables in order to increase throughput.

Below is a script I use when setting up web servers (Sol 10) which seems to

increase throughput - YMMV of course, and I wouldn't blindly apply the

script, take some time to work through it and identify if it would be useful

to you. Some features may not work on older Solaris versions, again YMMV.

http://docs.sun.com/app/docs?q=tunable+parameters+solaris is very useful.

#!/bin/sh
#
# S70stacktune
# v 1.22 25/2/2006 SamN
# Tune TCP stack on busy Solaris 10 boxes.

fill () {
awk '{leninput=length($($NF)); fill=63-leninput ; for (i=1; i< fill; i++)

fillchar=fillchar"." ; printf $($NF) fillchar}'
}

setparams () {
VAL=`/usr/sbin/ndd -set $1 $2 $3`
printf "Value of $1 $2 is: " | fill
printf " ${VAL} ($3)\n"
}

# Set the tcp time wait interval
printf "Set the tcp time wait interval\n"
setparams /dev/tcp tcp_time_wait_interval 60000

# Speed up the flushing of half-closed connection in state FIN_WAIT_2
printf "Flushing of half-closed connection in state FIN_WAIT_2\n"
setparams /dev/tcp tcp_fin_wait_2_flush_interval 67500ms

# Set the receive and transmit window sizes
printf "Set the receive and transmit window sizes\n"
setparams /dev/tcp tcp_xmit_hiwat 400000
setparams /dev/tcp tcp_recv_hiwat 400000

# Set number of half-open connections
printf "Set number of half-open connections\n"
setparams /dev/tcp tcp_conn_req_max_q0 102400

# Set number of simultaneous connections
printf "Set number of simultaneous connections\n"
setparams /dev/tcp tcp_conn_req_max_q 102400

# Set the maximum buffer size
printf "Set the maximum buffer size\n"
setparams /dev/tcp tcp_max_buf 4194304

# Set the tcp time wait interval
printf "Set the tcp time wait interval\n"
setparams /dev/tcp tcp_time_wait_interval 60000

# Set TCP connection abort interval
printf "Set TCP connection abort interval\n"
setparams /dev/tcp tcp_ip_abort_interval 60000

# Set congestion window size
printf "Set congestion window size\n"
setparams /dev/tcp tcp_cwnd_max 2097152

# Set TCP Keepalive Interval
printf "Set TCP Keepalive Interval\n"
setparams /dev/tcp tcp_keepalive_interval 60000

printf "Ok, tuning complete\n"
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: High TCP connect timeout rate questions
From: Kyle Tucker
References:
High TCP connect timeout rate questions
From: Kyle Tucker
Prev by Date: Re: select nic interface for outgoing IP messages
Next by Date: Re: Netstat Problem on solaris 5.8
Previous by thread: High TCP connect timeout rate questions
Next by thread: Re: High TCP connect timeout rate questions
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
---------
Re: High TCP connect timeout rate questions

-----------------------------------------------------------------------------

---

From: Sam N <sun@xxxxxxxxxxxxxx>
Date: Fri, 09 Jun 2006 06:46:24 GMT

-----------------------------------------------------------------------------

---
Kyle Tucker wrote:
Sam N <sun@xxxxxxxxxxxxxx> wrote:
Kyle Tucker wrote:
I have received this error on a port listened to by a Squid server. Is this

warning issued prior to reaching the point
where pending TCP requests would have been dropped? And if
that point is reached, would the counter for tcpListenDropQ0 shown in netstat

be greater than 0? Mine is at 0. Thanks.

It sounds like maybe the box is getting hammered - Is it really busy? There

are steps you can take to tune the stack on Solaris boxes, you can make

changes to set numerous variables in order to increase throughput.

I don't know how busy that particular box was at the time (no sar) and
it's at a customer site, but will watch it more closely in the future.

Below is a script I use when setting up web servers (Sol 10) which seems

[ script snipped ]

This looks very helpful.

http://docs.sun.com/app/docs?q=tunable+parameters+solaris is very useful.

Yes, I read a good amount on docs.sun.com and Google groups when this
occured. But I can't find a definitive answer to what actually occurs,
if anything, when the error is issued and what increments the above
counters in netstat. Any idea?

Thanks.

As far as I'm aware the alert is generated when the box thinks it's under a

SYN flood attack (and the unestablished connection queue becomes full).

/dev/tcp tcp_conn_req_max_q0 - unestablished connection queue (incomplete

handshake)
/dev/tcp tcp_conn_req_max_q - established connection queue

Are two parameters that you can tune in order to increase the queue sizes.

http://www.informit.com/articles/article.asp?p=101138&seqNum=5&rl=1 has more

info - it seems to be a pretty good article.

cheers

Sam

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: High TCP connect timeout rate questions
From: Kyle Tucker
References:
High TCP connect timeout rate questions
From: Kyle Tucker
Re: High TCP connect timeout rate questions
From: Sam N
Re: High TCP connect timeout rate questions
From: Kyle Tucker
Prev by Date: Re: shared library in 32 bit and 64 bit
Next by Date: Re: t2000, again (vs. V440s?)
Previous by thread: Re: High TCP connect timeout rate questions
Next by thread: Re: High TCP connect timeout rate questions
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: High TCP connect timeout rate questions
... Is this warning issued prior to reaching the point ... where pending TCP
requests would have been dropped? ... shown in netstat be greater than 0? ...
make changes to set numerous variables in order to increase throughput. ...

(comp.unix.solaris)
Re: High TCP connect timeout rate questions
... Is this warning issued prior to reaching the point ... where pending
TCP requests would have been dropped? ... shown in netstat be greater than 0?

...
(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

---------
Re: High TCP connect timeout rate questions

-----------------------------------------------------------------------------

---

From: kylet@xxxxxxxxxxxxx (Kyle Tucker)
Date: Fri, 9 Jun 2006 12:22:34 +0000 (UTC)

-----------------------------------------------------------------------------

---
In article <4P8ig.14582$qD.954@xxxxxxxxxxxxxxxxxxxx>,
Sam N <sun@xxxxxxxxxxxxxx> writes:

Sam N <sun@xxxxxxxxxxxxxx> wrote:

Kyle Tucker wrote:

I have received this error on a port listened to by a Squid
server. Is this warning issued prior to reaching the point
where pending TCP requests would have been dropped? And if
that point is reached, would the counter for tcpListenDropQ0
shown in netstat be greater than 0? Mine is at 0. Thanks.

It sounds like maybe the box is getting hammered - Is it really busy?
There are steps you can take to tune the stack on Solaris boxes, you can
make changes to set numerous variables in order to increase throughput.

I don't know how busy that particular box was at the time (no sar) and
it's at a customer site, but will watch it more closely in the future.

Below is a script I use when setting up web servers (Sol 10) which seems

[ script snipped ]

This looks very helpful.

http://docs.sun.com/app/docs?q=tunable+parameters+solaris is very useful.

As far as I'm aware the alert is generated when the box thinks it's
under a SYN flood attack (and the unestablished connection queue becomes
full).

Right, and then I'd have thought netstat would have showed
tcpListenDropQ0 > 0. The box had not been rebooted between
the message and when I ran netstat. Maybe I'll set up a test
box with a very low setting and see if I can get the error
generated and get tcpListenDropQ0 incremented. Thanks.

--
- Kyle
.

-----------------------------------------------------------------------------

---

References:
High TCP connect timeout rate questions
From: Kyle Tucker
Re: High TCP connect timeout rate questions
From: Sam N
Re: High TCP connect timeout rate questions
From: Kyle Tucker
Re: High TCP connect timeout rate questions
From: Sam N
Prev by Date: Re: Patching strategy for Solaris 10 Core Install
Next by Date: Re: Patching strategy for Solaris 10 Core Install
Previous by thread: Re: High TCP connect timeout rate questions
Next by thread: where is the "unsetenv" c function
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: High TCP connect timeout rate questions
... Is this warning issued prior to reaching the point ... where pending
TCP requests would have been dropped? ... shown in netstat be greater than 0?

...
(comp.unix.solaris)
Re: High TCP connect timeout rate questions
... Is this warning issued prior to reaching the point ... would the counter

for
tcpListenDropQ0 shown in netstat be greater than 0? ... There are steps you

can take to
tune the stack on Solaris boxes, you can make changes to set numerous

variables in order to increase
throughput. ... As far as I'm aware the alert is generated when the box

thinks it's under a SYN
flood attack (and the unestablished connection queue becomes full). ...

(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

3、Re: ufsdump very slow

-----------------------------------------------------------------------------

---

From: Dan Foster <usenet@xxxxxxxxxxx>
Date: Tue, 06 Jun 2006 15:49:54 -0500

-----------------------------------------------------------------------------

---
In article <4em6i7F1elij4U1@xxxxxxxxxxxxxx>, Michael Laajanen

<michael_laajanen@xxxxxxxxx> wrote:

I have noticed for some time(after changing tape device) that the new
tape (Ultrium 400GB compressed) is very little loaded(13%) during
ufsdump and that the disk that is beeing dumped is fully loaded(92%) but
with a very low transfer rate(4048), why is that so see below from
iostat during a level 0 dump.

The machine is a E250 and all disks in the machine are brand new Maxtor
10k 147GB disks exept for the root disk which is a original 26GB Seagate
10k drive.

extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 127.0 0.0 4064.1 0.0 0.1 0.0 1.1 0 13 rmt/0

4064.1 KB/sec divided by 127.0 writes/sec = 32 KB writes.

This is too small for LTO. LTO minimum for reasonable performance is 256
KB writes.

Performance is crawling because LTO drive has to slow down. You are not
feeding data to it fast enough for it to run at full speed.

Bottleneck here is how you are using ufsdump. Your tape drive and hard
drives are fast. :)

You might want to try ufsdump's 'b 512' argument to tell it to write
data to tape in 512 block chunks (256 KB).

ufsdump may limit it to 126 block chunks; not sure, so test and verify.

-Dan
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: ufsdump very slow
From: Michael Laajanen
References:
ufsdump very slow
From: Michael Laajanen
Prev by Date: Re: gethostbyname broken under Solaris 8
Next by Date: Re: gethostbyname broken under Solaris 8
Previous by thread: ufsdump very slow
Next by thread: Re: ufsdump very slow
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
RE: 306GB drives!
... 72GB scsi sca drives are avaialble for about $235 which works out to

about ...
>> storage disk, such that the signature can be read by a signal detector.

... >>
LTO tape drives actively doing the restore? ... (comp.os.vms)
Re: Need help on Tape,Optical disk...
... I do have extensive experience with both disk and tape in many ... Disk

drives,
on the other hand, ... They're sending 5 MB/s to a 60 MB/s drive, ...

(comp.arch.storage)
Re: W9xNT MCA Adapters - SPOCK-206
... IDs and find out which one is disk, CD, tape, scanner and so on, to be

... don't
recall offhand if CD-ROM drives trigger the lights, ... I seldom look at the

disk light
either. ... (comp.sys.ibm.ps2.hardware)
TAPE OR DISK STORAGE? MAYBE BOTH
... Tape or disk storage? ... The choice of enterprise storage systems ...
disk systems might suggest. ... drives to today's DLT and SuperDLT ...

(sci.chem)
TAPE OR DISK STORAGE? MAYBE BOTH
... Tape or disk storage? ... The choice of enterprise storage systems ...
disk systems might suggest. ... drives to today's DLT and SuperDLT ...

(sci.physics)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-----
Re: ufsdump very slow

-----------------------------------------------------------------------------

---

From: jayl-news@xxxxxxxxxxxxxx
Date: 9 Jun 2006 14:56:12 -0700

-----------------------------------------------------------------------------

---
Michael Laajanen wrote:

I don't know what is average, it around 480000 files which uses 35GB of
data.

Ummmm, well then, yes you *do* know the average, it's 73KB,
and yes, that qualifies as lots of little files. There is nothing
that will back this up as fast as you (apparently) want. Star
will be faster than gtar or ufsdump, but not 10X faster or
anything like that.

Search this group for "Backup tools too slow" and you'll get some
education.

service in spare time in our company, otherwise we would have MS
everywhere :)

But the backups would still be slow. :-)

2. If installa 5'th disk in the E250(which makes a total of 5 147 GB
disks), would some kind of SW raid help this?

Striping would help a little bit, but not 1.5X or anything like that.
It doesn't sound like you're ready for that in any case.

3. We where planning to more to S10 in december last year, but S10 was
delay(update) so we are still on S9, would S10 and ZFS help us?

You're getting 4MB/s IIRC, that's not bad. ZFS might help a little,
but ZFS provides flexibility and control, not so much improved
speed. ZFS cannot change average seek time on your drives.

What you *can* do is reduce the amount of time that live
data changes on the drive reduce the integrity of your
backup. The "backup window". You do this by either taking
a snapshot of the file system, and backing up the snapshot,
or doing an additional mirror of the file system, and do a
detach/backup/re-attach of that extra mirror.

ZFS makes the snapshot pretty easy. You can snapshot
your UFS filesystems, too, see fssnap. Last time I tried
incremental ufsdump on fssnap snapshots, it didn't handle
the timestamps correctly, but that was on Sol8, maybe
it's fixed now.

I reiterate, you should search this group, you'll get a lot
of good information.

-Jay-

-----------------------------------------------------------------------------

---

References:
ufsdump very slow
From: Michael Laajanen
Re: ufsdump very slow
From: Dan Foster
Re: ufsdump very slow
From: Michael Laajanen
Re: ufsdump very slow
From: Dan Foster
Re: ufsdump very slow
From: Michael Laajanen
Re: ufsdump very slow
From: Dan Foster
Re: ufsdump very slow
From: Michael Laajanen
Prev by Date: Re: How can I script shadow file modifications?
Next by Date: Re: How can I script shadow file modifications?
Previous by thread: Re: ufsdump very slow
Next by thread: Xorg on Solaris 8 SPARC
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Volume Shadow Copy on XP/2K3
... Most modern 3rd party backup software are VSS requestors - like, say,

Veritas
... which involve the data storage of aforementioned database software. ...

This
means that the snapshot can possibly contain non-integral in-transition ...

(comp.arch.storage)
Re: ZFS: filesystem approach
... >>with their own ZFS filesystem, to see how usable it would actually be?

...
so that if you wanted to use a traditional backup ... and the snapshots

appear in
a special ... (comp.unix.solaris)
Re: Looking for advice on rsync
... > they were asking about was using rsync or some other mechanism to ... I
explained that the backup can't really be ... never even notice the snapshot

...
We can mount the snapshot as another filesystem, ... (comp.os.linux.misc)
Re: batch file to delete files based on criteria
... >> I need to throw together a simple batch file to backup a file. ...
>> schedule this via Task Scheduler to run every hour or so. ... > Create a

snapshot
backup when a file changes. ... >!Snapshot script from inside Windows

Explorer.
... (microsoft.public.win2000.cmdprompt.admin)
Re: What are the advantages of LVM?
... enough that the backup is not self-consistant. ... now you can backup
the snapshot ... where you might have multiple virtual Linux machines running

(or
not ... (Ubuntu)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(04)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

4、Re: Networking, Zones, & Firewall Question w/ Solaris 10+

-----------------------------------------------------------------------------

---

From: aryzhov@xxxxxxxxx
Date: 6 Jun 2006 02:28:21 -0700

-----------------------------------------------------------------------------

---
Lion-O wrote:

It might be arguable that it's in fact *global* zone that routes the traffic
in this scenario. But look, global zone may have no valuable IP address

Then routing would be hardly possible since you /need/ an entry in your

routing
table to allow the global zone to access the non-global zone. So much for the
extra security...

I don't need an entry in routing table,
neither need I an IP address in the global zone,
in order to filter or RDR the packets for a non-global zone's subnet.

And when it comes to using routing daemons (in.routed / routed)
which can IMO be a very important part of a somewhat more advanced
router then your argument becomes plain out wrong.

Again, routing and firewalling are two very different activities.
Your statement is perfectly correct for routing but wrong
for packet filtering and NAT/RDR. Please give an example of
advanced firewall that needs a routing daemon.

There is no 'security advantage' to be gained here since the actual routing
will always be done on the global zone.

There's no routing in the global zone, but RDR tables only.
No routing daemons, either. Just a kernel module.
Totally different level in OS as well as in the IP stack.

The only real security advantage is to place network services
(http, mta, etc.) servers in your non-global zones.

I agree, but this has nothing to do with the topic.

Security advantage here is that administrator can avoid any
management traffic on subnets *and* any rule managenent activity
in virtual hosts being filtered, NATed or RDR'd.

Even though I use a separate Solaris 10 box as dedicated firewall,
I still consider it reasonable to configure the extranet and intranet
IP addresses in two separate non-global zones, and use a third
non-global zone (on a separate physical segment) for interactive
access.

From this management zone, I can ssh to a global zone via internal

virtual interface not visible from "production" zones.

Comparing with a traditional 3-segment firewall (intra, extra, mgmt),
I assign this 3 "functions" to 3 isolated machines, which gives me
some additional piece of mind.

Regards,
Andrei

-----------------------------------------------------------------------------

---

Prev by Date: Re: Faster backups - possibly Solaris pipe buffer size
Next by Date: Re: Size difference when copying a large file in solaris
Previous by thread: Solaris10 HW 06/06
Next by thread: gethostbyname broken under Solaris 8
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Networking, Zones, & Firewall Question w/ Solaris 10+
... Then routing would be hardly possible since you /need/ an entry in your

routing ...
table to allow the global zone to access the non-global zone. ... There is no

'security
advantage' to be gained here since the actual routing ... Because the

non-global zone
is in the routing table of the global zone. ... (comp.unix.solaris)
Re: Zones within a Zone?
... non-global zone can't influence the global zone (hint: ...

(comp.unix.solaris)
Zones and SMF issue
... I've got a package I produced. ... It works great in the global zone.
... So pkgadd running in the non-global zone cannot add that file. ...

(comp.unix.solaris)
Re: Configuring packages from zone to zone
... >samba in the global zone, ... >to my non-global zone by default,
but when I went to edit my smb.conf ... Why did you install the sunfreeware

version
and not the version ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

5、Re: how to find the total number of open file descriptors per application

-----------------------------------------------------------------------------

---

From: Robert Milkowski <rmilkowski-NO-SPAM@xxxxxxxx>
Date: Tue, 6 Jun 2006 04:30:59 +0000 (UTC)

-----------------------------------------------------------------------------

---
Rob <user@xxxxxxxx> wrote:

Hi,

How do I find the total number of file descriptors by application (just the
number not a list of open fd's)? I'm trying to monitor a server to see if
and when the rlims [in /etc/system (rlim_fd_max and rlim_fd_cur)] are
reached so that I can tune it.

This is Solaris 8 and it has lsof installed.

ls -l /proc/pid/fd/ | wc -l

however these wont't be only fds to files (network sockets too, etc).
Also 'man pfiles'.

--
Robert Milkowski
rmilkowskiADAADS@xxxxxxxx
http://milek.blogspot.com
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: how to find the total number of open file descriptors per application
From: Rob
References:
how to find the total number of open file descriptors per application
From: Rob
Prev by Date: how to find the total number of open file descriptors per

application
Next by Date: Re: how to find the total number of open file descriptors per

application
Previous by thread: how to find the total number of open file descriptors per

application
Next by thread: Re: how to find the total number of open file descriptors per

application
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: file descriptor to a command
... Rob wrote: ... > I am trying to learn file descriptors and thought
it might possibly ... > be useful to use them as a tmp file. ... If you want
to do something like this, you should use "named pipes". ...

(comp.unix.shell)
Re: Finding No. of open file descriptors in a process
... Note that in Solaris 10 pfiles includes the file name: ... carnot% pfiles
`pgrep vi` ... Current rlimit: 256 file descriptors ... (comp.unix.solaris)
Re: Finding Process ID of Runaway files
... gerg@panix.com (Greg Andrews) writes: ... output of pfiles on Solaris
7 and Solaris 8 look the same to me: ... Current rlimit: 256 file descriptors

...
(comp.sys.sun.admin)
Re: Finding Process ID of Runaway files
... gerg@panix.com (Greg Andrews) writes: ... output of pfiles on Solaris
7 and Solaris 8 look the same to me: ... Current rlimit: 256 file descriptors

...
(comp.unix.programmer)
Re: Finding Process ID of Runaway files
... gerg@panix.com (Greg Andrews) writes: ... output of pfiles on Solaris
7 and Solaris 8 look the same to me: ... Current rlimit: 256 file descriptors

...
(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(04)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-----

6、Re: New 1.4T SVM UFS fs does not have enough inodes

-----------------------------------------------------------------------------

---

From: Jim Prescott <jgp@xxxxxxxxxxxxxxxxxx>
Date: 5 Jun 2006 18:56:58 -0400

-----------------------------------------------------------------------------

---
In article <1149515332.749153.145950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Dougii <DouglasDenny@xxxxxxxxx> wrote:

How do I get more inodes in a multi-terabyte file UFS system?

Ask Sun to remove the arbitrary minimum of "nbpi=1048576" for EFI
filesystems. From what I've heard the limit is there because large
filesystems with lots of inodes take a long time to fsck. To me
a potential performance issue might justify a warning message and
a change in defaults, but not a hard limit.

If you can get by with .99T you can get as many inodes as you need.
Based on the df output you presented your current inode density is
around 211k so you could use "newfs -i 100000" on the .99T filesystem
to get a little extra usable space by making fewer inodes.

ZFS doesn't suffer from this limitation. I've also heard that with
the sources from OpenSolaris it is easy to bypass the EFI inode limit.
--
Jim Prescott - Computing and Networking Group jgp@xxxxxxxxxxxxxxxxxx
School of Engineering and Applied Sciences, University of Rochester, NY
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Peter Eriksson
References:
New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
Prev by Date: Re: gdb for Solaris 10
Next by Date: Re: Faster backups - possibly Solaris pipe buffer size
Previous by thread: New 1.4T SVM UFS fs does not have enough inodes
Next by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: performance of filesystem xattrs with Samba4
... filesystems at: ... Andreas to add support to ext3 for large inodes.
... and only lost a very small amount when xattrs were not ... clearly how

much
we are paying for journaling. ... (Linux-Kernel)
Re: 1 Gigabyte SD card "full" after 467 MB [SOLVED]
... filesystems, including ext3, where it's possible to run out of inodes ...

In
practice most filesystems have huge numbers of inodes, ... the real problem

must have been
inodes, directory entries, etc. ... just have nautilus (or konqueror or any

other
application you want to ... (Ubuntu)
Re: All filesystems change to read-only after a while with kernel 2.6.6-1.435
... > of my filesystems (ext3) change to read-only. ... > about inodes
being members of orphan nodes. ... > that and the subsequent reboot works

fine. ...
(Fedora)
Re: New 1.4T SVM UFS fs does not have enough inodes
... Dougii wrote: ... filesystems with lots of inodes take a long time to
fsck. ... to get a little extra usable space by making fewer inodes. ... One

problem
with this many inodes is that fsck will use a *LOT* of RAM ...

(comp.unix.solaris)
Re: Can I use a round bobbin redirects in a shell ?
... names do not have permissions. ... > A name (struct dirent) is a pointer
to an inode. ... "File serial number", nothing about inodes. ... >> Where
is it stated in POSIX that all filesystems must use inodes? ...

(uk.comp.os.linux)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: New 1.4T SVM UFS fs does not have enough inodes

-----------------------------------------------------------------------------

---

From: Peter Eriksson <peter@xxxxxxxxxx>
Date: Tue, 6 Jun 2006 05:17:55 +0000 (UTC)

-----------------------------------------------------------------------------

---
Jim Prescott <jgp@xxxxxxxxxxxxxxxxxx> writes:

In article <1149515332.749153.145950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Dougii <DouglasDenny@xxxxxxxxx> wrote:

How do I get more inodes in a multi-terabyte file UFS system?

ZFS doesn't suffer from this limitation. I've also heard that with
the sources from OpenSolaris it is easy to bypass the EFI inode limit.

Yup. It's really simple. You can get a premodified version from:

ftp://ftp.ifm.liu.se/pub/unix/mkfs-sol10-ifm.tar

One problem with this many inodes is that fsck will use a *LOT* of RAM
to fsck the filesystem and this might start swapping which easily will
cause things to take virtually forever...

You definitiely want to use logging on these filesystems.

(I use this in a fat filesystem that I use as an rsync target from a
number of other servers - so if everything fails I can just newfs
the filesystem and re-rsync everything).

- Peter
--
--
Peter Eriksson <peter@xxxxxxxxxx> Phone: +46 13 28 2786
Computer Systems Manager/BOFH Cell/GSM: +46 705 18 2786
Physics Department, Linköping University Room: Building F, F203
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
References:
New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Jim Prescott
Prev by Date: Re: ntp in cluster mode
Next by Date: Re: Instal curl
Previous by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Next by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

much
we are paying for journaling. ... (Linux-Kernel)
Re: 20TB Storage System
... > you may not want to create filesystems over 3TB if you want fsck to ...
I don't know if that's using the default newfs settings ... > (which would

create an
insane number of inodes), ... (freebsd-hackers)
Re: 20TB Storage System
... > you may not want to create filesystems over 3TB if you want fsck to ...
I don't know if that's using the default newfs settings ... > (which would

create an
insane number of inodes), ... (freebsd-performance)
Re: 20TB Storage System
... > you may not want to create filesystems over 3TB if you want fsck to ...
I don't know if that's using the default newfs settings ... > (which would

create an
insane number of inodes), ... (freebsd-questions)
Re: 1 Gigabyte SD card "full" after 467 MB [SOLVED]
... filesystems, including ext3, where it's possible to run out of inodes ...

In
practice most filesystems have huge numbers of inodes, ... the real problem

must have been
inodes, directory entries, etc. ... just have nautilus (or konqueror or any

other
application you want to ... (Ubuntu)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
----
Re: New 1.4T SVM UFS fs does not have enough inodes

-----------------------------------------------------------------------------

---

From: "Dougii" <DouglasDenny@xxxxxxxxx>
Date: 7 Jun 2006 03:58:05 -0700

-----------------------------------------------------------------------------

---
Thanks to Jim and Peter on this. I opened a ticket with Sun and indeed
they told me the limitation for TB filesystems is 1024*1024 / TB. I
will ask sun to remove this limit as part of my ticket, but I will not
hold my breath.

I am going to look at ZFS again, but it is still not clear to me if it
is in general release, i.e. supported under Solaris 10. (3/05 or 1/06).
Maybe it is time to start testing ZFS on some QC systems.

-Doug

Peter Eriksson wrote:

Jim Prescott <jgp@xxxxxxxxxxxxxxxxxx> writes:

In article <1149515332.749153.145950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Dougii <DouglasDenny@xxxxxxxxx> wrote:

How do I get more inodes in a multi-terabyte file UFS system?

ZFS doesn't suffer from this limitation. I've also heard that with
the sources from OpenSolaris it is easy to bypass the EFI inode limit.

Yup. It's really simple. You can get a premodified version from:

ftp://ftp.ifm.liu.se/pub/unix/mkfs-sol10-ifm.tar

One problem with this many inodes is that fsck will use a *LOT* of RAM
to fsck the filesystem and this might start swapping which easily will
cause things to take virtually forever...

You definitiely want to use logging on these filesystems.

(I use this in a fat filesystem that I use as an rsync target from a
number of other servers - so if everything fails I can just newfs
the filesystem and re-rsync everything).

- Peter
--
--
Peter Eriksson <peter@xxxxxxxxxx> Phone: +46 13 28 2786
Computer Systems Manager/BOFH Cell/GSM: +46 705 18 2786
Physics Department, Linköping University Room: Building F, F203

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Darren Dunham
References:
New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Jim Prescott
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Peter Eriksson
Prev by Date: Re: Configuration of Solaris OS
Next by Date: Re: Configuration of Solaris OS
Previous by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Next by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

much
we are paying for journaling. ... (Linux-Kernel)
Re: 1 Gigabyte SD card "full" after 467 MB [SOLVED]
... filesystems, including ext3, where it's possible to run out of inodes ...

In
practice most filesystems have huge numbers of inodes, ... the real problem

must have been
inodes, directory entries, etc. ... just have nautilus (or konqueror or any

problem
with this many inodes is that fsck will use a *LOT* of RAM ...

(uk.comp.os.linux)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06
-----
Re: New 1.4T SVM UFS fs does not have enough inodes

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Wed, 07 Jun 2006 20:17:37 GMT

-----------------------------------------------------------------------------

---
Dougii <DouglasDenny@xxxxxxxxx> wrote:

No. It will be included in the 6/06 release.

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

References:
New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Jim Prescott
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Peter Eriksson
Re: New 1.4T SVM UFS fs does not have enough inodes
From: Dougii
Prev by Date: multiple processes of mysql/httpd running...
Next by Date: Re: Can't see where is the problem
Previous by thread: Re: New 1.4T SVM UFS fs does not have enough inodes
Next by thread: Users cannot login after changing file permissions
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-06

7、Re: ntp in cluster mode

-----------------------------------------------------------------------------

---

From: Logan Shaw <lshaw-usenet@xxxxxxxxxxxxx>
Date: Tue, 06 Jun 2006 04:59:02 GMT

-----------------------------------------------------------------------------

---
Przem wrote:
I need advice when configuring ntp on cluster.

For now it uses /etc/inet/ntp.conf.cluster and server 127.127.1.0 as time

server.
But because of that this cluster at this moment 12 minutes faster than my

whole environment.

I haven't set up a cluster, although I have used ntp a fair bit.

My advice would be that you need to get your clock calibrated
on every server before you set them up to try to run independently.
The most straightforward way I can think of to do that is to run
ntpd against an external server for a week or so, then let it build
an ntp.drift file that it can use to compensate for the natural
drift of the system's clock.

How can I configure it to use external ntp server?
As I can see, /etc/init.d/xntpd.cluster stopes when it find

/etc/inet/ntp.conf, so I assume I should still use

/etc/inet/ntp.conf.cluster.

Can I simply change field "server 127.127.1.0" to i.e. "server

ntp.server.my"? Will nodes be still synchronized than?

If you have Internet connectivity available, the easiest thing would
be to just synchronize your servers to external servers. If you
really want to ensure that your local machines stay in sync with
each other, you can solve that in one of at least two ways:

(1) synchronize each internal server with enough external servers
that the chances of one of them being far off is really small,
just because it would require too many failures (of external
servers) for that to be likely to happen, or
(2) synchronize one server with external servers, then synchronize
your other servers to it; then your internal servers will not
all be on equal footing because one will be the master, but
they should still stay in sync for a while if you lose internet
connectivity for a day or two.

In general, keep in mind that just setting up a cluster of servers
trying to sync to each other will not give you reliable and
accurate timekeeping. Computers make OK clocks but not good ones,
and in order to keep in sync, you need to ultimately be drawing
your information from a good clock (either from a public ntp server
or from a hardware clock, like a GPS receiver).

My actulaf conf files are (both same):
server 127.127.1.0
peer clusternode1-priv prefer
peer clusternode2-priv
driftfile /var/ntp/ntp.drift
filegen peerstats file peerstats type day enable
filegen loopstats file loopstats type day enable
filegen clockstats file clockstats type day enable

how should they look after I switch to external ntp source?

Just look at what they suggest doing at http://www.pool.ntp.org/ .
If you are in Europe, you should be able to put

server 0.europe.pool.ntp.org
server 1.europe.pool.ntp.org
server 2.europe.pool.ntp.org

and remove all the "peer" lines, and that should probably do it
for you.

- Logan
.

-----------------------------------------------------------------------------

---

References:
ntp in cluster mode
From: Przem
Prev by Date: Re: how to find the total number of open file descriptors per

application
Next by Date: Re: New 1.4T SVM UFS fs does not have enough inodes
Previous by thread: ntp in cluster mode
Next by thread: New 1.4T SVM UFS fs does not have enough inodes
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Couple of clustering questions
... vmware gsx 2 and build your envisioned cluster environment on that to

test. ...
NT 4.0 servers and are looking to upgrade to 2003 server and want to explore

... that I
want to fail over, but about software that does not need redundancy. ...

nodes at the same
time that I need to divide the SAN into multiple disks. ...

(microsoft.public.windows.server.clustering)
Re: Exch 2003 SP2 - applied on one node, but cant move resources
... resources to Node2, the failover did not complete because 'system

attendant' ...
Virtual Exchange server and failover occurred normally again upon taking ...

cluster
resources oline. ... (microsoft.public.exchange.admin)
Re: Couple of clustering questions
... I have had several issues related to VMWare GSX ... >vmware gsx 2 and

build
your envisioned cluster ... server but does not ... Do all clustered services
that can fail over use ... (microsoft.public.windows.server.clustering)
Re: Basic clustering Q
... Whats wrong with a bit of server consolidation? ... You can never have

enough
DC's to support that cluster! ... >> Hot spare hard disk to provide extra
FT across all disk configs. ... it's about support and having a working

configuration ...
(microsoft.public.windows.server.clustering)
Re: Best Configuration for a 3 Node SQL 2000 Cluster on Windows 2003?
... > I'm not sure if I'm setting up SQL right... ... > cluster, but
I want to make sure we don't shoot ourselves in the foot. ... dedicated SQL

server should
have minimal need for a paging ... During install time you can select which

cluster nodes
to install ... (microsoft.public.sqlserver.clustering)

--------------------------------------------------------------

随机文章：

0604月7，8，9日 2006-06-13

0603月19，20，21，22，23，24日 2006-04-18

0602月22，23，24，25，26，27，28日 2006-03-14

0511月7，8，9日 2005-12-25

0506月4，5，6日 2005-07-30

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0605月25，26，27，28，29，30，31日

mmmmn — Wed, 26 Jul 2006 23:17:49 +0800

1、Re: T1000 and internal disks

-----------------------------------------------------------------------------

---

From: "Mark" <mark.round@xxxxxxxxx>
Date: 31 May 2006 07:05:19 -0700

-----------------------------------------------------------------------------

---

Hi all

To settle an argument here, can the T1000 support 2 internal drives, or
only one? I've had conflicting advice from Sun representatives.

We've just got hold of one of these units, with some more to come.
There is only physically room inside for one SATA drive - although if
you follow the SATA cable from the drive back to the motherboard, it
appears there is is another similar connector right next to it. I'm
guessing that you could probably hook another SATA drive up to it and
it would work - but you'd have to get the power to it from somewhere
(there's only one drive power connector as well), and find some way of
shoe-horning it in there with the other drive, and face all the
additional problems of airflow and cooling etc., not to mention voiding
your chance of ever getting support for it!

So, the answer is pretty much "no", although I suppose in theory you
could probably do it somehow. You have got a single PCI-E slot though,
so if you need extra storage, you could always use a PCI-E fibre HBA
(there's one that's intended for the T2000 that fits and I see they've
now qualified it for use in the T1000) and hook up to more storage that
way...

-Mark

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: T1000 and internal disks
From: Sam N
References:
T1000 and internal disks
From: Sam N
Prev by Date: Re: T1000 and internal disks
Next by Date: Re: T1000 and internal disks
Previous by thread: Re: T1000 and internal disks
Next by thread: Re: T1000 and internal disks
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
... I have tried disabling readahead and DMA on both drives ... My

motherboard is
an ABIT NF7-S2G nFORCE2 with the following IDE ... # ACPI Support ... #

Infrared-port
device drivers ... (Linux-Kernel)
Re: "Safely remove hardware" icon is missinbg
... mainstream support so I'm not sure if these hotfixes are still available

...
their drives after all, and they suggested ... indicate the unplug option or

icon.
... and then double-click the appropriate icon. ...

(microsoft.public.win2000.new_user)
Re: Latest on Word 2003 Crashing Display.
... Internet Support site of the vendors. ... :> also tried the support

forums
in Microsoft without success. ... :> 1) Reformatting both Drives and

reinstalling
everything, ... font will stop typing and the blue font will start. ...

(microsoft.public.word.application.errors)
Atapicam problems
... drives show up in /dev again and 'camcontrol devlist' lists them. ...

atapicam0:
... options SOFTUPDATES # Enable FFS soft updates support ...

(freebsd-stable)
Re: Updated SATA support patches for Intel ICH & Promise cards
... |> This is only HW support and does not address other ata/FreeBSD issues.

...
| I havn't looked into hotswapping SATA much yet, ... SATA drives since you

have
to read the registers and do reset etc. ... this especially for things like

Intel when
they just expose the SATA ... (freebsd-current)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05

2、Re: Running X based application from web server

-----------------------------------------------------------------------------

---

From: Rainer Beushausen <surfnasi@xxxxxx>
Date: Wed, 31 May 2006 15:38:41 +0200

-----------------------------------------------------------------------------

---
Dave (from the UK) wrote:

Martin Paul wrote:

In comp.unix.solaris "Dave (from the UK)"
<see-my-signature@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:

I'm trying to run an X application (mathematica) via a web server
which is running on a Sun Ultra 60. I do *not* want to see the
graphical output (which is a 3D graph), but save it to a file. The
application supports saving files to about n million different formats.

While I haven't used it personally, I think Xvfb can be used for this
purpose - see the man page or google.

mp.

Thanks, that does (did) sound useful. I like the bit in the man page:

... and providing an unobtrusive way to run applications
that don't really need an X server but insist on having one
anyway.

But it still does not work. I know the program is trying to access
display :11, so I tried

main-webserver /usr/local/apache2/logs # Xvfb :11
Warning: There is no XDISPLAY information for display 11.
Server is using XDISPLAY information for display 0.

But when I try to run the application, the following error is generated:

X11 connection rejected because of wrong authentication.
X11 connection rejected because of wrong authentication.
X11 connection rejected because of wrong authentication.
X11 connection rejected because of wrong authentication.
X11 connection rejected because of wrong authentication.
X11 connection rejected because of wrong authentication.

How it is supposed to authenticate?

# Xvfb -help
use: X [:] [option]
-a # mouse acceleration (pixels)
-ac disable access control restrictions
-audit int set audit trail level
-auth file select authorization file
etc etc

I've tried the -ac option, but that did not help. I'm not sure what the

-auth file

is supposed to do. I seems I may have authentication problems, but how
does one get around it? Put something in 'file' ? What something?

Any ideas anyone?

Would suggest to use Xnest (usr/openwin/bin/Xnest)

Rainer
.

-----------------------------------------------------------------------------

---

References:
Re: Running X based application from web server
From: Martin Paul
Re: Running X based application from web server
From: Dave (from the UK)
Prev by Date: Re: Running X based application from web server
Next by Date: Re: T1000 and internal disks
Previous by thread: Re: Running X based application from web server
Next by thread: Change ipaddress on Solaris 10 box
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: HELP Connection error on Release mode
... "Off" Always display detailed ASP.NET error information. ... This section

sets
the authentication policies of the application. ... Set trace enabled="true"

to
enable application trace logging. ...

随机文章：

0604月1，2，3，4，5，6日 2006-06-12

0507月28，29，30，31日 2005-09-08

0507月7，8，9日 2005-08-28

0505月7，8，9日 2005-07-03

0504月22，23，24，25，26日 2005-06-20

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0605月19，20，21，21，23，24日

mmmmn — Wed, 26 Jul 2006 00:24:33 +0800

1、Re: Geforce 6600 driver hassles on Solaris 10 1/06

-----------------------------------------------------------------------------

---

From: "Dave (from the UK)" <see-my-signature@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Thu, 25 May 2006 11:21:00 +0100

-----------------------------------------------------------------------------

---

Any suggestions on the best driver to use? I assume download one from
the Nvida site?

I'm a newbie to the OpenSolaris world and had pretty much the same exact
experience.

I've not been using OpenSolaris - I had tried Solaris Express, but are having

problems installing that. I think it might be a DVD creation problem now, as

I can get it to boot from a CD properly, but only half-boot from a DVD.

During the install the VESA mode was usable but, alas, there
was no permanent configuration for X installed and the subsequent reboot
after successful install meant no X.

It seems this is a problem, which might be easily fixed. One assumes if a

driver works once, it should work again. But perhaps the issue is during

install the driver is 32-bit, but then needs to be 64-bit - for my Opteron at

least. I'm not sure what CPU you have.

I found that the xorgconfig text based
script produced very close to a functional xorg.conf if the same VESA mode
was used. This at least made X usable but unable to get out of VESA mode as
the nv driver simply wouldn't work.

I'm currently using JDS and found that after installing the new drivers from

the Nvidia site, and and following instructions from here:

http://www.phoronix.com/scan.php?page=article&item=406&num=1

I got it working. Briefly, that needed.

# /usr/X11/bin/scanpci

where it produced an output similar to this for me.

pci bus 0x0004 cardnum 0x00 function 0x00: vendor 0x10de device 0x0141
nVidia Corporation GeForce 6600 GT

Then I needed to run

# update_drv -a -i '"pci10de,0141"' nvidia.

and do a reconfigure boot.

Should probably mention my card is a PNY 6800GT AGP card on an old KD7A
KT400A chipset mobo. The AGP port driver in FreeBSD did not support this
AGP port

I'm not using AGP, but PCI Express

I realise it must be hard with all the different hardware around, but this

really need to be sorted out whereby failures like this occur. Two things

happened that I think should be avoided.

1) The install picks the wrong driver. There should at that point be a stage

where it tests the driver, tells the user it is testing, and asks them to

report if it works or not. Currently the default is to use it without doing a

test. Sure you can test, but most will not.

2) After selectomg a driver that allows the install to go ahead, and spent

some time answering all the questions about networking etc, it is annoying to

find you can't see anything sensible. If the driver has worked during the

install, I'm not sure why it should not work once the isntall is complete.

Perhaps though, a dirrferent driver must be used, sicne one is 32-bit and the

other 64-bit. I guess that is harder to check, unless the installer can be

forced to work in 64-bit too, so at least one is testing the same driver.

--
Dave K MCSE.

MCSE = Minefield Consultant and Solitaire Expert.

Please note my email address changes periodically to avoid spam.
It is always of the form: month-year@domain. Hitting reply will work
for a couple of months only. Later set it manually.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Geforce 6600 driver hassles on Solaris 10 1/06
From: Alan Coopersmith
Prev by Date: Re: cdrecord: What mode (TAO, DAO, SAO ..etc ) to write a

bootable DVD?
Next by Date: How to detect the MIME type with a commad?
Previous by thread: cdrecord: What mode (TAO, DAO, SAO ..etc ) to write a

bootable DVD?
Next by thread: Re: Geforce 6600 driver hassles on Solaris 10 1/06
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: XP HOME BOOT FAILURE
... you remove the driver for the graphics card, ... >> If you run a repair
install as I initially described, ... A repair install should bypass that but

it will
... >>>> on, follow the screens. ...

(microsoft.public.windowsxp.accessibility)
Re: Modem Install Problems
... PCI: PCI BIOS revision 2.10 entry at 0xfb9c0, ... ACPI tables contain no

PCI
IRQ routing entries ... Serial driver version 5.05c with HUB-6 MANY_PORTS ...
install: cannot stat `536ep.o': No such file or directory ...

(comp.os.linux.setup)
RE: Word 2003 "not enough memory or disk space to run word"
... Another possibility is that you have a network printer driver that is not

...
install new printer drivers on your hard disk, the only fix is to make sure

... whether
or not you decide to continue to run the NAV plug-in for Office ?...

(microsoft.public.word.application.errors)
Re: No floppy device in FC5
... FC5 does, although with the caveat that a lot of folks seem to be having

...
and ATI graphics driver problems. ... I had a lot of problems installing FC4,

...
Going the Linux way, not the MS (you need to re-install the ... (Fedora)
Re: SilentInstall
... For a driver to install silently, ... #I060 Set selected driver.
... #-166 Device install function: DIF_REGISTER_COINSTALLERS. ...

(microsoft.public.development.device.drivers)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05

--------

2、Re: Network interface configuration

-----------------------------------------------------------------------------

---

From: Sam N <sun@xxxxxxxxxxxxxx>
Date: Thu, 25 May 2006 08:06:59 GMT

-----------------------------------------------------------------------------

---
googler wrote:
Hi all,

Hi there

I am trying to configure the network interface in a Solaris 10
workstation (I had skipped this step during installation). This is the
first time I'm doing this. I did read up some articles, man page and
older posts in this newsgroup, but still a little confused about the
exact steps to follow.

I understand that I have to configure the interface by the following
command line:
# ifconfig e1000g0 xx.yy.zz.ww broadcast xx.yy.zz.255 netmask
255.255.255.0 plumb up

1. To run the above command, I need the IP address xx.yy.zz.ww of the
host. I'm trying to get it through DHCP. I followed the instructions at
this webpage: http://www.rite-group.com/consulting/solaris_dhcp.html
(create the files /etc/hostname.e1000g0 and /etc/dhcp.e1000g0 manually
and reboot). After reboot, I was expecting that the IP address for this
interface will be shown. But this is what I got:
# ifconfig -a
lo0: flags=2001000849 mtu
8232 index 1
inet 127.0.0.1 netmask ff000000
e1000g0: flags=1004802 mtu 1500 index 2
inet 0.0.0.0 netmask 0
ether 0:3:ba:f1:3f:fe

It looks like the interface hasn't come up - try # ifconfig e1000g0 up

At this point I figured that I could have done the same as above using
ifconfig command itself, so wanted to see if it makes any difference
(that is, if it can assign the IP address). So my next command was:
# ifconfig e1000g0 auto-dhcp primary wait 120
ifconfig: e1000g0: interface currently has a pending command (try
later)

Confused, I wanted to see if the following command would work.
# ifconfig e1000g0 plumb
ifconfig: SIOCSLIFNAME for ip: e1000g0: already exists

The interface is already plumbed in, you can't plumb it in twice. To remove

it and start afresh, try # ifconfig e1000g0 unplumb

So how do I get the IP address?

The easiest way to start this process again would be to run "sys-unconfig"

and follow the prompts.

2. How do I know the name of the network interface? It took me some
time (reading older posts and checking /etc/drivers_aliases) to figure
out that it is e1000g0 in my case. Is there an easier way?

Try a prtconf - you should get the driver name and instance from this. In the

case below, we can see 4 interfaces, ipge0, ipge1, ipge2, ipge3.

# prtconf -D |grep network
network, instance #0 (driver name: ipge)
network, instance #1 (driver name: ipge)
network, instance #2 (driver name: ipge)
network, instance #3 (driver name: ipge)

Thanks for your help.

No worries.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Network interface configuration
From: googler
References:
Network interface configuration
From: googler
Prev by Date: Re: v40z with Opteron vs v240
Next by Date: Re: Solaris 10 Reverts back to half duplex when restarted
Previous by thread: Re: Network interface configuration
Next by thread: Re: Network interface configuration
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: [SLE] Wireless Internet at public hotspots -- update
... I need another command to force the card ... What I see from ifconfig
is that eth1 will show an IP ... Now, all my previous network ... showing me

an
IP that wasn't really assigned by DHCP? ... (SuSE)
Re: Theory Question
... You are missing my points (or perhaps I typed too fast to make them ...

Crist
supplied the ifconfig scenario, I just followed up on it, ... With the

correct supporting
measures in place, obscuring network topology is ... (FreeBSD-Security)
Re: [SLE] 10.0: no net, now what, contd
... > When you're in Linux open a console shell issue the command ... > the
ifconfig command, and place it at the beginning of that line, left ... >

hardware
problem. ... >> YaST from the desktop, and how to add its icon there.) ...
(SuSE)
Re: ifconfig command shows only loopback device.
... after running modprobe e1000 ifconfig returns eth0. ... ifconfig -a does
not show eth0. ... configure for e1000 network module to be loaded at every

boot?. ...
(RedHat)
Re: Questions with configuring multiple NICs
... I can ifconfig sis0, the onboard ethernet, ... I can't reach anything off

the
local network. ... > cards configured correctly with ifconfig. ... >
the modules and ifconfig'ing only the interface I need at that point in ...

(freebsd-questions)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05
--------

3、Re: Which FCAL to buy for 3510 array?

-----------------------------------------------------------------------------

---

From: dol@xxxxxxxxxxxxxx (Fredrik Lundholm)
Date: Wed, 24 May 2006 11:06:03 +0000 (UTC)

-----------------------------------------------------------------------------

---
In article <1148435313.236627.260760@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
<zcorce@xxxxxxx> wrote:

While reviewing the options from the Sun Sore there are many FCAL PCI
options available:

Are there definite advantages of one particular PCI card over another?

Note that the BUS interfaces differ,
so what card is best depends on your server. (PCI-express, PCI-X)
There are two varieties of each card, qlogic and emulex.
I would choose the qlogic, since it is more proven.

Can the 3510 array support 4 GB speeds in direct connect configurations
or are the 4GB cards for SAN implementations?

You could use the 4GB card directly with the 3510, it will use the 2GB speed.
Since the 4GB is the same price as the 2GB, take the 4GB, IF YOU HAVE SERVERS
WITH SPARE 3.3VOLT PCI SLOTS. The V240 only has one, the V490 only two,
V1280 only one etc.

If you use pci-express, take the fastest available. (T1000, T2000, X2100)

/Regards
Fredrik

--
Fredrik Lundholm
dol @ ce.chalmers.se

-----------------------------------------------------------------------------

---

References:
Which FCAL to buy for 3510 array?
From: zcorce
Prev by Date: Re: May 23 10:50:05 mwsbit04 pseudo: [ID 129642 kern.info]

pseudo-device: devinfo0
Next by Date: Re: redirecting cronjob errors messages
Previous by thread: Which FCAL to buy for 3510 array?
Next by thread: Solaris 10 on Sun T1000
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
RE: OT: Sparc not dead yet
... When the server reboots, it again starts taking ... I've implimented Load
Broker and supplimented it with LAN/IP failover ... Since replacing a card in

a set
might only take 5-10 ... I thought in terms of NICs we said LAN ...

(comp.os.vms)
RE: My Server Died
... That got my active directory back but the DHCP was still ... I wanted to

use
VPN and needed to add another NIC card. ... 3COM said to install older

versions
of the card first, ... > suitable domain controller and when I tell it this

server is
the last domain ... (microsoft.public.win2000.advanced_server)
RE: WinTV-Go and XDMCP
... I have a old ATI-all-in-wonder as my graphics card. ... :-) I have a

collection
of old PCs that I am setting up as a little server farm. ... Claremont pings

amito
and amito ... >> connection, is there? ... (Fedora)
Re: Long log in time
... you activate the second network card and attach the router to the second

...
to ensure the LAN nic is at the top (I usually rename the nics LAN and ...

>From your
IP configs I believe your router is 10.0.0.1 and server is ... The only thing

you could
do with the 2nd card is to enable the SBS ...

(microsoft.public.windows.server.sbs)
Re: My Server Died
... Printer Sharing for Microsoft Networks installed. ... > I had this server
working properly on my LAN for a long time. ... I wanted to use VPN and

needed to add another
NIC card. ... 3COM said to install older versions of the card first, ...
(microsoft.public.win2000.advanced_server)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05

4、Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays

-----------------------------------------------------------------------------

---

From: "Stuart" <stuart.savill@xxxxxxxxxxxx>
Date: 23 May 2006 06:04:26 -0700

-----------------------------------------------------------------------------

---
Hi all,

quick question - working on a automated solaris 10 build for our
customers, and they support a number of different HBA's and Storage
arrays - I wonder if anyone has come across a good reference point for
suggested Kernel Tunables for storage within solaris

for some of detail - HBA's are a mix of the usual suspects from Emulex
and Qlogic
Storage arrays are (again) usual suspects from EMC / HDS etc.

any words of wisdom would be greatfully recieved.

Many thanks,

Stuart

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays
From: bl_antczak
Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays
From: Dexthor
Prev by Date: Re: May 23 10:50:05 mwsbit04 pseudo: [ID 129642 kern.info]

pseudo-device: devinfo0
Next by Date: Re: Strange controller numbering with Solaris 10
Previous by thread: May 23 10:50:05 mwsbit04 pseudo: [ID 129642 kern.info]

pseudo-device: devinfo0
Next by thread: Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage

Arrays
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: E4500 compatibility with new FC/SCSI arrays
... I am assuming that you are using Qlogic adapters on the host to connect

...
I dont know what storage stack you are running. ... With Qlogic adapters and

SFS
patches, ... > of the new storage arrays, and possibly our old boxes too. ...
(comp.unix.solaris)
Re: Hard disk drive as big as 1600Gi
... largest arrays I have used are over 250TB. ... things like screen I/O.
... You have the latency times involved in seeking, ... It'd be better if you

said
what you actually want to do with the storage. ...

(alt.comp.periphs.mainboard.asus)
Re Profiling Lisp code
... This is not a quest for the fastest cross product.. ... >quest for the

better
storage method. ... than consider using arrays, because the array is the very

fundamntal
... depending on the appication, ... (comp.lang.lisp)
Re: Returning arrays and Objects in C++
... Arrays are so objects in C++, although it is not possible to return ...

An object
is a region of storage. ... The term object type refers to the type with ...
While bare arrays cannot be passed to, or returned from, functions by ...

(comp.lang.cpp)
E4500 compatibility with new FC/SCSI arrays
... of the new storage arrays, and possibly our old boxes too. ... Our

current version
of Veritas is no longer supported and we've been ... The other solution would

be a 3310
as a) SCSI arrays have got faster, ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05
-----
Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays

-----------------------------------------------------------------------------

---

From: "Dexthor" <gmolakal@xxxxxxxxx>
Date: 23 May 2006 06:14:49 -0700

-----------------------------------------------------------------------------

---
In the past I had to touch the following when dealing with Gigabit
HBAs.

maxphys
sd:max_throttle
vol_maxio
bufhwm/vx_bufhwm (to avoid buffer caching)

Emulex, I think, had a good write up on Solaris kernel settings for
optimal through put in their adapter manuals. If you are using a
volume manager, you have to follow their best pratices so that your
configuration is optimal to your workload.

HTH
-Dexthor.

-----------------------------------------------------------------------------

---

References:
Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays
From: Stuart
Prev by Date: Re: Strange controller numbering with Solaris 10
Next by Date: Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage

Arrays
Previous by thread: Solaris 10 / Kernel Tunables / Storage HBA's / Storage

Arrays
Next by thread: Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage

Arrays
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05
--------
Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays

-----------------------------------------------------------------------------

---

From: "bl_antczak" <bl.antczak@xxxxxxxxx>
Date: 23 May 2006 06:15:07 -0700

-----------------------------------------------------------------------------

---
Hello,

You should find some usefull PDF's according Qlogic on:

http://support.qlogic.com/support/drivers_software.asp

You must find model of HBA - there will be some info about tunable
parameters according kind of storage

and for Emulex:

http://www.emulex.com/ts/index.html

or if You have access to powerlink.emc.com - try to look on their
knowledge base.

But generaly will be very hard to compare one value for various HBA's
and storage. I think, that You should check proposals for all and
compare if they are not totally different :-)

Regards,
Blazej Antczak

Stuart napisal(a):

Hi all,

for some of detail - HBA's are a mix of the usual suspects from Emulex
and Qlogic
Storage arrays are (again) usual suspects from EMC / HDS etc.

any words of wisdom would be greatfully recieved.

Many thanks,

Stuart

-----------------------------------------------------------------------------

---

References:
Solaris 10 / Kernel Tunables / Storage HBA's / Storage Arrays
From: Stuart
Prev by Date: Re: Solaris 10 / Kernel Tunables / Storage HBA's / Storage

Arrays
Next by Date: Re: backup issue
Previous by thread: Re: Solaris 10 / Kernel Tunables / Storage HBA's /

Storage Arrays
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05

5、Re: "Normal" home directory and UID for useradd ?

-----------------------------------------------------------------------------

---

From: Keith Thompson <kst-u@xxxxxxx>
Date: Fri, 19 May 2006 23:26:57 GMT

-----------------------------------------------------------------------------

---
"Daniel Rock" <v200620@xxxxxxxxxxx> writes:

Richard B. Gilbert <rgilbert88@xxxxxxxxxxx> wrote:

It would be nice if the system just assigned the UID the next available
number but no Unix system I know of does this.

Solaris searches for the highest UID already assigned and uses the next
one.

I believe that there is
a convention that "normal" UIDs start at 100 but AFAIK it is not required.

You should really skip the first 100 UIDs. Although not all of them are
used currently you may run into trouble in the future when more and more
"system users" are introduced:

With sendmail 8.12ff you got user smmsp (uid 25)
With JDS you got user gdm (uid 50)
With Solaris 10 you got user webservd (uid 80)
etc.

The useradd program manpage explicitly warn against using UIDs < 100:

(UIDs from 0-99 are reserved for possible use in future
applications.)

Some systems allocate user accounts starting at UID 500. Linux
typically does this, although the "useradd" man page doesn't mention
this; it does say that "Values between 0 and 99 are typically reserved
for system accounts." I think what's going on here is that the system
installation routine, which prompts you for the name of the first user
account, explicitly sets its UID to 500; thereafter, other accounts
will start at 501.

If you want users to have consistent UIDs across different types of
systems (useful if you use NFS), you might consider assigning UIDs for
user accounts starting at 500. On the other hand, if you explicitly
assign UIDs, starting at 100 seems to work, at least across Solaris,
Linux, and a few other Unix-like OSs.

--
Keith Thompson (The_Other_Keith) kst-u@xxxxxxx <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.
.

-----------------------------------------------------------------------------

---

References:
"Normal" home directory and UID for useradd ?
From: Catherina Svensson
Re: "Normal" home directory and UID for useradd ?
From: Richard B. Gilbert
Re: "Normal" home directory and UID for useradd ?
From: Daniel Rock
Prev by Date: Re: Solaris 10: Cannot lu system
Next by Date: Re: PRoblems with DNS Access
Previous by thread: Re: "Normal" home directory and UID for useradd ?
Next by thread: Re: "Normal" home directory and UID for useradd ?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: root equivalent user
... > I don't think it applies to Linux but on some Unixes, small UIDs ...

RBAC
is not historical, though. ... Solaris 8 IIRC, FCS date of Feb 2000, and is

present in
Solaris 9 and ... (comp.os.linux.setup)
Re: "Normal" home directory and UID for useradd ?
... number but no Unix system I know of does this. ... Solaris searches for
the highest UID already assigned and uses the next ... You should really skip

the first
100 UIDs. ... (comp.unix.solaris)
tar from Solaris 8 to Solaris 10?
... On a Solaris 10 platform, can I extract a tarfile of a directory

hierarchy ...
UIDs and GIDs are identical on ... (comp.unix.solaris)
Re: tar from Solaris 8 to Solaris 10?
... > On a Solaris 10 platform, can I extract a tarfile of a directory

hierarchy
... UIDs and GIDs are identical on ... (comp.unix.solaris)
Re: is there a way to copy over just a user account?
... It was because of this limit that our accounts system was originally ...

designed
to reuse uids. ... Ian Gregory ... (comp.sys.mac.system)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-05

随机文章：

0511月25，26，27日 2006-01-16

0510月10，11，12日 2005-11-21

0508月19，20，21日 2005-10-03

0507月22，23，24日 2005-09-07

0507月4，5，6日 2005-08-23

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0604月19，20，21，22，23，24日

mmmmn — Mon, 26 Jun 2006 00:12:40 +0800

1、Re: mpathd question

-----------------------------------------------------------------------------

---

From: Scott Howard <scott@xxxxxxxxxxxxxxxxx>
Date: Thu, 27 Apr 2006 06:24:56 -0500

-----------------------------------------------------------------------------

---
cconnell_1@xxxxxxxxx wrote:

The in.mpathd daemon sends the ICMP echo request probes to on-link
routers. If no routers are available, it sends the probes to
neighboring hosts. Thus, for network failure detection and repair,
there must be at least one neighbor on each link that responds to ICMP
echo request probes.

In our topology there are 2 cisco routers with vrrp gateways (one is
not the default gateway).

And we need mpathd to base its up/down status on these, but how does
mpathd know these are gateways? Since it looks like it just chooses any
ip on the network to ping to?

If there is a default router, it will use that as the one and only test
target.

If there is no default router it will use multicast to find up to 5 other
hosts (not routers[1]) on the network, and use them as test target.
Under Solaris 8 and (currently) 9 this can cause issues as if all of the
test hosts go away then it will fail the link and not attempt to find any
new test target.

Under Solaris 10 (and 9, once we get the backport out - hopefully in a
month or so) it will add and remove hosts from it's list on the fly
as they go away/appear.

[1] - it uses a multicast "all hosts" broadcast to find targets. Some
routers will respond to this multicast, some will not.

Scott.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: mpathd question
From: cconnell_1
References:
mpathd question
From: cconnell_1
Prev by Date: Troubleshooting fiber channel GBICs
Next by Date: Re: Troubleshooting fiber channel GBICs
Previous by thread: mpathd question
Next by thread: Re: mpathd question
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Routing
... I Think in first your Netmask from your eth0 is wrong. ... Net Adress and

253
and 254 can you tahe for the Hosts. ... It only reaches the opposite side

network
... My network expert says that there is no issue in the routers. ...

(comp.os.linux.networking)
Re: redundant network interface question
... there'll be always hosts running on that network. ... It tries routers
first. ... (comp.unix.solaris)
Re: VPN Firewall
... >>>that model will only support 10 concurrent VPN tunnels. ... How
would I run two routers in ... shouldn't route ip. ... probably want to run
an igp that your hosts can understand (usually this ...

(comp.security.firewalls)
Re: Virtual Network testing
... For i386 machine emulation you can use vmware ... hosts and routers.
... Not sure how you to emulate Switches and HUB. ... (Pen-Test)
Re: Mix hosts and routers on a subnet?
... I've always avoided having routers exchange routing updates on a subnet

that carries
hosts. ... Back then a design objective was to keep routing segments as pure

as
possible because you had many protocols like IPX that were very "chatty" and

LAN bandwidth was at a premium
because of the collisions issue. ... (comp.dcom.sys.cisco)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

2、Re: snoop question

-----------------------------------------------------------------------------

---

From: James Carlson <james.d.carlson@xxxxxxx>
Date: 28 Apr 2006 09:11:51 -0400

-----------------------------------------------------------------------------

---
"Stuart" <bigdakine@xxxxxxx> writes:

When running snoop -V -tr, I noticed that the ID field which seems to
normally increment by one, sometimes skips a few numbers, and I was
wondering if that was indicative of a problem.

Solaris currently tracks ID numbers on the IRE cache entries for a
peer. This means that, for a given remote address, the numbers used
will normally increase linearly (sequentially). The exception is when
the IRE cache entry gets flushed, which happens either when it hasn't
been used in a while (no traffic to that peer) or when there's a
kernel forwarding table (routing) change that forces cache entries to
be recomputed.

--
James Carlson, KISS Network <james.d.carlson@xxxxxxx>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: snoop question
From: Stuart
References:
snoop question
From: Stuart
Re: snoop question
From: James Carlson
Re: snoop question
From: Stuart
Prev by Date: Re: Solaris 10 x86
Next by Date: Re: How to get the time difference in a script
Previous by thread: Re: snoop question
Next by thread: Re: snoop question
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(04)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
Re: snoop question

-----------------------------------------------------------------------------

---

From: James Carlson <james.d.carlson@xxxxxxx>
Date: 27 Apr 2006 08:15:34 -0400

-----------------------------------------------------------------------------

---
"Stuart" <bigdakine@xxxxxxx> writes:

when runing snoop -V

what is the meaning of the ID= field?

Is that simply the packet number or is it something else?

It's the IP Identification field. See RFC 791.

--
James Carlson, KISS Network <james.d.carlson@xxxxxxx>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: snoop question
From: Stuart
References:
snoop question
From: Stuart
Prev by Date: Re: ARP broadcast arrives, but MAC not changed
Next by Date: Re: Real solaris 9 newbie GUI question
Previous by thread: snoop question
Next by thread: Re: snoop question
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

2、Re: Setting a "default" project(1M) for a user ?

-----------------------------------------------------------------------------

---

From: Robert Milkowski <rmilkowski-NO-SPAM@xxxxxxxx>
Date: Wed, 26 Apr 2006 23:02:20 +0000 (UTC)

-----------------------------------------------------------------------------

---
Roger P. Johnson <roger0080@xxxxxxxxxxxx> wrote:

I am looking to create a project on Solaris 10 for a database server like

everyone else does.

I found a real simple how-to here:

http://forum.sun.com/jive/thread.jspa?threadID=74218
Briefly the author states to create the project like this:

# projadd -U oracle -K "project.max-shm-memory=(priv,2048MB,deny)"

user.oracle

He also states that "- because the name was of the form "user.username" it

becomes the oracle user's default project."

Where is this documented about the user. and default projects? I cannot

find any references to a default project for a user in the projadd(1M),

projmod(1M) or in the "System Administration Guide: Solaris

Containers-Resource Management and Solaris Zones (817-1592-10)" manual where

it walks you through projects.

All I can find concerning a dot in the project name is that "the period is

reserved for projects with special meaning to the operating system".

Hmmm. Should one be creating projects of the form user.??

How is this project he creates different from creating the project as below:

# projadd -U oracle -K "project.max-shm-memory=(priv,2048MB,deny)" projoracle

????

As I understand it, in the first case, when the user "oracle" logs in, the

oracle project is in effect (since it is their default)? After logging in

they can start their oracle database server up.

And in the second projadd example, when the user oracle logs in, would they

have to issue the command (?) below:

$ newtask -v -p oracleproj

to get or inherit the oracleproj resource settings and then be able to start

their oracle database server up? Yes/No ???

TIA

Exactly. Check it with 'id -p'.

To read about user.X look at: 'man -s 4 project' 'man getdefaultproj'.

--
Robert Milkowski
rmilkowskiADFASD@xxxxxxxx
http://milek.blogspot.com
.

-----------------------------------------------------------------------------

---

References:
Setting a "default" project(1M) for a user ?
From: Roger P. Johnson
Prev by Date: Setting a "default" project(1M) for a user ?
Next by Date: Re: SunRay 1g vs 2/2FS
Previous by thread: Setting a "default" project(1M) for a user ?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
US-CERT Technical Cyber Security Alert TA05-292A -- Oracle Products Contain

Multiple
... Oracle Products Contain Multiple Vulnerabilities ... Oracle released
a Critical Patch Update in October 2005. ... have the Oracle Database Server

installed).
... (comp.security.announce)
US-CERT Technical Cyber Security Alert TA05-292A
... > Oracle Products Contain Multiple Vulnerabilities ... > Oracle
released a Critical Patch Update in October 2005. ... > that do not have the

Oracle
Database Server installed). ... (comp.databases.oracle.misc)
Patch available for multiple critical flaws in Oracle
... Researchers at NGSSoftware have discovered multiple critical

vulnerabilities ...
in Oracle Database Server and Oracle Application Server. ... The

vulnerabilities range
from buffer overflow issues, PL/SQL Injection, ... (NT-Bugtraq)
Patch available for multiple critical flaws in Oracle
... Researchers at NGSSoftware have discovered multiple critical

vulnerabilities ...
in Oracle Database Server and Oracle Application Server. ... The

vulnerabilities range
from buffer overflow issues, PL/SQL Injection, ... (Bugtraq)
[VulnWatch] Patch available for multiple critical flaws in Oracle
... Researchers at NGSSoftware have discovered multiple critical

vulnerabilities ...
in Oracle Database Server and Oracle Application Server. ... The

vulnerabilities range
from buffer overflow issues, PL/SQL Injection, ... (VulnWatch)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

3、Re: Track process

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Wed, 26 Apr 2006 17:42:43 GMT

-----------------------------------------------------------------------------

---
Dolphin <Dolphin@xxxxxxxxx> wrote:

Hi,

Below is the iostat (iostat -xnp) information from one of the SUN servers
(E4500 running solaris 8 connecting to T3 via SAN switch).

extended device statistics

r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device

41.3 27.0 3111.2 958.8 2.5 5.8 3*** 84.5 1 47 c10t3d0

When I run prstat -a, almost all processes are sleeping. How to track the
process that contributed to high disk I/O?

With Solaris 10 you could run dtrace. Even though you might not want to
run it in production, if you could have a test box with similar load it
might be instructive.

For Solaris 8, you're limited to kernel I/O tracing. You can use 'prex'
to trace for a period of time, then the 'tnf' utilities to dump the
output into a file. A quick perl script can then find the heaviest I/O
via PID. Not quick and easy, but very possible.

See:
http://www.sunmanagers.org/pipermail/summaries/2004-January/004951.html
http://www.itworld.com/Comp/2402/UIR961001perf/

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Track process
From: Dolphin
References:
Track process
From: Dolphin
Prev by Date: Re: McNealy quits
Next by Date: OpenMOSIX equivalent for Solaris?
Previous by thread: Track process
Next by thread: Re: Track process
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
SUMMARY: Q: [Oracle 8i e450 UFS] Solaris8->9 upgrade: Performance Boost?-

Real World Experience ?
... feature, "Concurrent DirectIO", which ?apparently? ... Database Access".
... data on a character device [Raw device, Quick I/O, Direct I/O, etc..] ...

SOlaris
Platform_ references directio specifically. ... (SunManagers)
Re: Solaris versions and performance information (+slow Oracle + high WAIT

I/O)
... 90's) they still seem valid even for Solaris 8. ... We did have an I/O

problem
in the sense that queries related to ... CPU minf mjf xcal intr ithr csw icsw

migr smtx srw
syscl usr sys ... (comp.unix.solaris)
Re: Problems with very slow disks on a SUN A1000 disk-array!
... Block sizes and caching in I/O ... You can't really mess with Solaris'

internal
settings but dd can be ... DeeDee, don't press that button! ...

(comp.unix.solaris)
Re: Solaris versions and performance information (+slow Oracle + high WAIT

I/O)
... > In searching for information about performance and tuning for Solaris

...
This is due to the way WAIT I/O is measured. ... what you regard as excessive

user time
and io wait. ... (comp.unix.solaris)
Solaris versions and performance information (+slow Oracle + high WAIT I/O)
... performance problem with an Oracle database on Solaris 8. ... In

searching for
information about performance and tuning for Solaris ... This is due to the

way
WAIT I/O is measured. ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
-----
Re: Track process

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Thu, 27 Apr 2006 16:10:43 GMT

-----------------------------------------------------------------------------

---
Dolphin <Dolphin@xxxxxxxxx> wrote:

I ran prex followed by tnfdump after reading the URL. There is a process
with high delta time (5.7 sec) compared to the rest with device: 88003 and
flag: 524569.

man tnfdump does not show much information. Do you know where to find
information about how to decode the device: 88003 and flag:524569?

I don't know specifically, but I belive it is an encoding of the minor
number of the device. So a device with some combination of the numbers
'1', '87', or '343', and '195' would be my best guess.

If you don't have too many, you can go discovering. Use a particular
process to write somewhere. Then trace and look for I/O by that PID.

The server has 6 CPU but the tnfdump output show 9 under CPU column. Any
pointer?

CPUs are not always numbered sequentially. Compare with the output of
'psrinfo'.

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

References:
Track process
From: Dolphin
Re: Track process
From: Darren Dunham
Re: Track process
From: Dolphin
Prev by Date: Re: Solaris NDIS Wrapper Toolkit and usb wifi dongle
Next by Date: Solaris x86 Browser
Previous by thread: Re: Track process
Next by thread: Second Hard Drive Partition Confusion
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
RE: [Fwd: The modem chicken and egg question? Newbie Alert!]
... compare yours to his. ... hint: a space. ... you must always put
a space after the name of the command and sometimes ... in between each flag.

...
(RedHat)
Re: Filter address book on PR_EMS_AB_HOME_MDB
... PR_EMS_AB_HOME_MDB and a string to compare with. ... I use 0 for the flag
... Dmitry Streblechenko wrote: ... > OutlookSpy - Outlook, CDO ...
(microsoft.public.win32.programmer.messaging)
Re: QI and MQ Coder: First real-life experiences
... Either coder can be changed to decode from input or output size ...

normalize
such figure to N*logconvention or to other coders which ... That way you can

compare
your output size with the exact N*logvalue ... opposed to declaring the

smallest value
seen). ... (comp.compression)
Re: Byte?
... > say i need to decode the binary byte recieved on MSComm1.Input so that

the ...
> value of each bit determines a flag setting or what ever and i have the ...

If And
&H80) Then ' Check if high bit is set ... > what can i compare it to? ...
(microsoft.public.vb.general.discussion)
Re: Compare 2 PAS files
... I prefer a freeware one. ... it will even show decode and ...
show differences between binary DFM files. ... I think that Beyond Compare

has a
trial version. ... (borland.public.delphi.thirdpartytools.general)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

3、Re: Solaris 10 kstat output of multicore machine

-----------------------------------------------------------------------------

---

From: "Daniel Rock" <v200617@xxxxxxxxxxx>
Date: Mon, 24 Apr 2006 16:52:59 +0000 (UTC)

-----------------------------------------------------------------------------

---
Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> wrote:

Hi,

can anybody post a kstat output of a multicore machine running Solaris
10? I need to know if the enumeration of cpu_stat/instance is consistent
with cpu/sys/instance. intrstat, cpustat, and cpuinfo look consistent,
but I really would like to know for sure.

A complete output would be rather large, but basically:

module = cpu
instance = processor_id
name = intrstat/sys/vm

% kstat -m cpu | grep -v '^ '
module: cpu instance: 0
name: intrstat class: misc

module: cpu instance: 0
name: sys class: misc

module: cpu instance: 0
name: vm class: misc

module: cpu instance: 1
name: intrstat class: misc

module: cpu instance: 1
name: sys class: misc

module: cpu instance: 1
name: vm class: misc

For "kstat -m cpu -n sys" exist a second structure with almost the same
content. It is just named differently:

module = cpu_stat
instance = processor_id
name = cpu_stat%d (%d = processor_id)

% kstat -m cpu_stat|grep -v '^ '
module: cpu_stat instance: 0
name: cpu_stat0 class: misc

module: cpu_stat instance: 1
name: cpu_stat1 class: misc

Unfortunately the instances of cpu_stat are not enumerated sequentally
on multicore cpus which makes collecting the information a bit tricky,
when one consideres that a cpu could be turned off during this process.

Does anybody know, if there is an upper limit of instances. kstat_lookup
takes an int, which is a pretty long way to go, if a cpu got turned off.

Just iterate through all cpu_stat%d entries. Or call
sysconf(_SC_CPUID_MAX)
to know up to which instance number you have to iterate. Remember there
may be holes in the processor numbering, esp. on SPARC.

--
Daniel
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Solaris 10 kstat output of multicore machine
From: Thomas Maier-Komor
Re: Solaris 10 kstat output of multicore machine
From: Thomas Maier-Komor
References:
Solaris 10 kstat output of multicore machine
From: Thomas Maier-Komor
Prev by Date: Re: Getting GUI for Solaris!
Next by Date: Re: Compiler error using C++ V6 Update 2
Previous by thread: Solaris 10 kstat output of multicore machine
Next by thread: Re: Solaris 10 kstat output of multicore machine
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Solaris 10 kstat output of multicore machine
... can anybody post a kstat output of a multicore machine running Solaris

... I
need to know if the enumeration of cpu_stat/instance is consistent ... takes

an
int, which is a pretty long way to go, if a cpu got turned off. ...

(comp.unix.solaris)
Re: Solaris 10 kstat output of multicore machine
... can anybody post a kstat output of a multicore machine running Solaris

... I
need to know if the enumeration of cpu_stat/instance is consistent ... I know

- and even
more important: cpu/sys is Solaris 10 and later only, ... when one consideres

that a cpu
could be turned off during this process. ... (comp.unix.solaris)
Re: [PATCH v3]Export cpu topology by sysfs
... > 2) Set consistent default values for the 4 attributes. ... > If one
architecture wants to support this feature, ... Just include itself, if the

cpu
doesn't support ... is on powerpc) is useful regardless of the cpu's online

state. ...
(Linux-Kernel)
100% cpu usage for LSASS.EXE on DC intermittently, consistent inte
... On my AD User Domain a DC which has PDC & GC roles spiked CPU at

consistent ...
it was LSASS.EXE which was consuming CPU ... patterns (no unusual spikes) but

the
issue moved to another DC in the User ... SAMBA users correctly configured

who appearently
don't cause this problem ... (microsoft.public.win2000.networking)
RE: [PATCH v3]Export cpu topology by sysfs
... >>> 2) Set consistent default values for the 4 attributes. ... >>>
If one architecture wants to support this feature, ... If the cpu has only 1

core,
we could call it single-core, so it's reasonable to use 0 as its core id. ...

>>is on powerpc)
is useful regardless of the cpu's online state. ... (Linux-Kernel)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(07)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----

4、Re: kstat & sysinfo

-----------------------------------------------------------------------------

---

From: Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx>
Date: Mon, 24 Apr 2006 10:44:15 +0200

-----------------------------------------------------------------------------

---
Thomas Maier-Komor wrote:

Hi,

does anybody know how one can access the cpu_stats_t structure (declared
in sys/sysinfo.h) via the kstat interface? cpu_stat_t is accessible via
"cpu_stat". But in the kstat output there is no cpu_stats and the values
of cpu_stats_t are shown under cpu_stat.

BTW: this is on Solaris 9.

Any ideas or pointers?

TIA,
Tom

I mixed it up a little bit. The information of sys/sysinfo.h's
cpu_stats_t is included in the kstat output, but the structure is only
defined in Solaris 10. Solaris 9' sys/sysinfo.h misses this structure,
but the information must be there, because kstat can show it.

Any idea how to get it?

TIA,
Tom
.

-----------------------------------------------------------------------------

---

References:
kstat & sysinfo
From: Thomas Maier-Komor
Prev by Date: kstat & sysinfo
Next by Date: libmtsk ERROR while loading a so
Previous by thread: kstat & sysinfo
Next by thread: Re: kstat & sysinfo
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Problems with elxl on s10_72 x86
... 100Mbs bandwidth. ... Here is the relevant output of kstat -c
net ... This has been like this for quite a while (all Solaris versions I

have ever ...
(comp.unix.solaris)
Re: networking with a pc question
... > I am using solaris 5.7. ... kstat doesn't seem to exist. ...
Oh, my, an oldie. ... 'le' does 10Mbps half-duplex only, ...

(comp.unix.solaris)
Re: How can I measure the network load ?
... > So I look for a network monitor that I can use ... For solaris,
try the perl script on ... simply run kstat with the right arguments and

parse the
output). ... (comp.unix.misc)
Re: How can I measure the network load ?
... > So I look for a network monitor that I can use ... For solaris,
try the perl script on ... simply run kstat with the right arguments and

parse the
output). ... (comp.unix.questions)
Re: How can I measure the network load ?
... > So I look for a network monitor that I can use ... For solaris,
try the perl script on ... simply run kstat with the right arguments and

parse the
output). ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
---
Re: kstat & sysinfo

-----------------------------------------------------------------------------

---

From: "Daniel Rock" <v200617@xxxxxxxxxxx>
Date: Mon, 24 Apr 2006 09:12:24 +0000 (UTC)

-----------------------------------------------------------------------------

---
Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> wrote:

Hi,

cpu_stats are new in Solaris 10. It is basically the same as cpu_stat, but
the values are 64 bits wide instead of 32 bits. The user/sys/wait/idle
times are also converted from clock ticks to nanoseconds.

kstat -m cpu_stat

--
Daniel
.

-----------------------------------------------------------------------------

---

References:
kstat & sysinfo
From: Thomas Maier-Komor
Prev by Date: Re: Adventures with ILOM on an X4100
Next by Date: Re: problem during configure of apache
Previous by thread: Re: kstat & sysinfo
Next by thread: libmtsk ERROR while loading a so
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(09)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

5、Re: need help setting up my solaris 10 machine as a dns server

-----------------------------------------------------------------------------

---

From: "stroller" <smcbutler@xxxxxxxxxxx>
Date: 23 Apr 2006 14:31:16 -0700

-----------------------------------------------------------------------------

---

hi, i know something about this but i'm confused on a couple of points

1) where is the global list of services available to me?

2) how do i know which services are managed by this and which by the
old /etc/init.d mechanism?
eg: apache seems to be controlled by the old way, but when i try to
enable it i get the message "svcs" Pattern 'enable' doesn't match any
instances

also, when i tried the command you mentioned "svcadm enable
/network/dns/server:default" and i run "svcs | grep dns" it just comes
back with:

online 13:37:03 svc:/network/dns/client:default
offline 14:16:00 svc:/network/dns/server:default

i can't get the dns server to enable at all, maybe this isn't properly
setup (i copied over all my named and zone files from my solaris 8
machine). is there a way to tell?

thx for your help!

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: need help setting up my solaris 10 machine as a dns server
From: Gary Mills
References:
need help setting up my solaris 10 machine as a dns server
From: stroller
Re: need help setting up my solaris 10 machine as a dns server
From: Sandz
Prev by Date: Re: Help reviving a sol8 box. Please help...
Next by Date: Re: need help setting up my solaris 10 machine as a dns server
Previous by thread: Re: need help setting up my solaris 10 machine as a dns

server
Next by thread: Re: need help setting up my solaris 10 machine as a dns

server
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Page Cannot Be Displayed
... Turn all Computers OFF and Unplug the Router from the Power Cord ... Ping
statistics for 192.168.0.1: ... Step1: Then Try this Start The RUN Command

and type
in the following: ... and look for the DNS Server is there is any thing in

front of it
like Enable ... (microsoft.public.windows.inetexplorer.ie6.browser)
Re: Page Cannot Be Displayed
... Turn all Computers OFF and Unplug the Router from the Power Cord ... Ping
statistics for 192.168.0.1: ... Step1: Then Try this Start The RUN Command

and type
in the following: ... and look for the DNS Server is there is any thing in

front of it
like Enable ... (microsoft.public.windows.inetexplorer.ie6.browser)
Re: OO Design question about batch job
... > fit the OPs problem space is how Oliver is proposing the lack of use of

...
A Model - View pattern degenerates ... MVC is like a fancier version of

Observer;
... > one command or strategy class for each type of file. ...

(comp.lang.java.programmer)
Re: NIS 2005 prevents net use w: ...
... I permitted SMB in my trusted zone. ... command would just hang. ...
The second subnet uses the address range 192.168.1.x. ... and a DNS server.

...
(microsoft.public.security)
Re: Page Cannot Be Displayed
... You did not answer my Q about if the Router is Firewall Built one and if

you ...
Ping statistics for 192.168.0.1: ... Step1: Then Try this Start The RUN

Command
and type in the following: ... and look for the DNS Server is there is any

thing in front
of it like Enable ... (microsoft.public.windows.inetexplorer.ie6.browser)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: need help setting up my solaris 10 machine as a dns server

-----------------------------------------------------------------------------

---

From: "stroller" <smcbutler@xxxxxxxxxxx>
Date: 24 Apr 2006 14:28:48 -0700

-----------------------------------------------------------------------------

---

i think i figured it out..

to start the dns server i was trying svcadm -restart and svcadm -start
but these had no effect.

i tried svcadm disable and then svcadm enable and it came up
successfully

so svcadm doesn't behave the same as the old init.d scripts..

anyways, thx very much for your help!!

-----------------------------------------------------------------------------

---

References:
need help setting up my solaris 10 machine as a dns server
From: stroller
Re: need help setting up my solaris 10 machine as a dns server
From: Sandz
Re: need help setting up my solaris 10 machine as a dns server
From: stroller
Re: need help setting up my solaris 10 machine as a dns server
From: Gary Mills
Re: need help setting up my solaris 10 machine as a dns server
From: stroller
Re: need help setting up my solaris 10 machine as a dns server
From: stroller
Prev by Date: Re: How to tell what model Ultra 1E workstation? 140E/170/200
Next by Date: Re: How to tell what model Ultra 1E workstation? 140E/170/200
Previous by thread: Re: need help setting up my solaris 10 machine as a dns

server
Next by thread: install and config of SNMP agent as mon-root
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

随机文章：

0605月25，26，27，28，29，30，31日 2006-07-26

0604月10，11，12，13，14，15日 2006-06-15

0509月13，14，15日 2005-10-18

0503月17，18，19日 2005-06-02

0502月16，17，18日 2005-05-12

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0604月10，11，12，13，14，15日

mmmmn — Thu, 15 Jun 2006 02:42:38 +0800

1、Re: How to check the number of "user processes" in the system?

-----------------------------------------------------------------------------

---

From: "tunla" <lars.tunkrans@xxxxxxxxxxxx>
Date: 17 Apr 2006 14:14:30 -0700

-----------------------------------------------------------------------------

---

jacksuyu@xxxxxxxxx wrote:

HOw to check the number of "user processes" in solaris?

Also how to increase or decrease it?

Thanks.

Solaris 10:

If you want to count number of processes in the Interactive class
casued by interactively logged on users.

ps -elc | grep IA | wc -l

Default number of user procs is 29.995
youll need a 100 MB Ram per user
each JDS user will run an average of 35 processes

29995/35 = 857 users *100 MB = 86 GB Ram

therefore the smallest machine where this can become a concern
is a SunFire 2900 with 12 ultrasparc IV CPU and 96 GB ram.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: How to check the number of "user processes" in the system?
From: Crazy Eyes
References:
How to check the number of "user processes" in the system?
From: jacksuyu
Prev by Date: Re: how to customize sun keyboard
Next by Date: Re: 32 bit or 64 bit?
Previous by thread: How to check the number of "user processes" in the

system?
Next by thread: Re: How to check the number of "user processes" in the

system?
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
---------
Re: How to check the number of "user processes" in the system?

-----------------------------------------------------------------------------

---

From: "Crazy Eyes" <phusnikn@xxxxxxxxx>
Date: Tue, 18 Apr 2006 00:21:52 GMT

-----------------------------------------------------------------------------

---
This is wrong, as most of the memory allocated would be pointing to the same
phsyical mapped pages hense shared.

"tunla" <lars.tunkrans@xxxxxxxxxxxx> wrote in message
news:1145308470.924594.271290@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

jacksuyu@xxxxxxxxx wrote:

HOw to check the number of "user processes" in solaris?

Also how to increase or decrease it?

Thanks.

Solaris 10:

If you want to count number of processes in the Interactive class
casued by interactively logged on users.

ps -elc | grep IA | wc -l

Default number of user procs is 29.995
youll need a 100 MB Ram per user
each JDS user will run an average of 35 processes

29995/35 = 857 users *100 MB = 86 GB Ram

therefore the smallest machine where this can become a concern
is a SunFire 2900 with 12 ultrasparc IV CPU and 96 GB ram.

-----------------------------------------------------------------------------

---

References:
How to check the number of "user processes" in the system?
From: jacksuyu
Re: How to check the number of "user processes" in the system?
From: tunla
Prev by Date: Re: Warning at boot time with Solaris 9
Next by Date: Re: smdc connections on my x2100 server
Previous by thread: Re: How to check the number of "user processes" in the

system?
Next by thread: can't su to root
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

2、Re: how to customize sun keyboard

-----------------------------------------------------------------------------

---

From: "Heinz M黮ler" <onkelheinz@xxxxxxxxxxxx>
Date: Thu, 20 Apr 2006 19:59:06 +0200

-----------------------------------------------------------------------------

---
Thanks, that works fine!!!

heinz

"Stefaan A Eeckels" <hoendech@xxxxxx> schrieb im Newsbeitrag
news:20060418104412.e5304b66.hoendech@xxxxxxxxx
On Mon, 17 Apr 2006 18:15:15 +0200
"Heinz Müller" <onkelheinz@xxxxxxxxxxxx> wrote:

is there a way to customize F1, F2, F3 etc. with commands like ps -ef?

I'm getting tired to write commands like "ps -ef | more" or "ls -ltr".
(Yes, I know I can specify aliases).

SUN Solaris 8 with SUN Keyboard
SUN Ultra 5

There is when you're using bash or tcsh.

To map F6 to "ps -ef" in tcsh, use the bindkey built-in:

$ bindkey -c ^[[17~ "ps -ef"

To enter the F6 key sequence, press "control-v F6". Obviously,
this can be placed in your start-up file (.cshrc).

In bash, the same is achieved through the "bind" built-in. Use "read"
to find the key sequence emitted by F6

$ read
^[[17~

The ^[ is the ESC character, which in bash is represented by
\e (backslash e), so to map F6 onto "ps -ef" you'd type:

$ bind '"\e[17~": "ps -ef^M"'

(Press "control-v control-m" to enter the Return.)

On start-up, bash reads the ~/.inputrc file to map keys or character
sequences to shell commands.

Take care,

--
Stefaan A Eeckels
--
Governments are like babies: digestive tracts with a big appetite at
one end and no sense of responsibility at the other. The better run
ones from time to time get clean diapers...

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: how to customize sun keyboard
From: Stefaan A Eeckels
References:
how to customize sun keyboard
From: Heinz M黮ler
Re: how to customize sun keyboard
From: Stefaan A Eeckels
Prev by Date: partioning slices on primary disk for mirroring
Next by Date: Re: partioning slices on primary disk for mirroring
Previous by thread: Re: how to customize sun keyboard
Next by thread: Re: how to customize sun keyboard
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: how to customize sun keyboard
... SUN Solaris 8 with SUN Keyboard ... There is when you're using bash or
tcsh. ... To enter the F6 key sequence, ... The ^[is the ESC character, which
in bash is represented by ... (comp.unix.solaris)
Re: ASCII hexadecimal number to character?
... On Sun, 14 Dec 2003 at 00:16 GMT, St閜hanie Vanhove wrote: ... Only if
bash is your shell. ... (comp.unix.shell)
Re: ASCII hexadecimal number to character?
... On Sun, 14 Dec 2003 at 00:16 GMT, St閜hanie Vanhove wrote: ... Only if
bash is your shell. ... (comp.unix.questions)
Re: Off-topic Sun evolution, was Re: Fetchmail misbehaviour?
... I think it's OpenSSH based but with some SUN modifications. ... I know

that
in Solaris 8 bash is 2.0.x, tcsh is ... (comp.os.linux.security)
Re: Machine restart without information
... SauDoi wrote: ... I have connected from my PC to Sun via console cable.
... and turning off the pc. ... It's also possible to change the key sequence
on the Sun that will ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

3、Re: How long is a piece of string?

-----------------------------------------------------------------------------

---

From: KJ <here@xxxxxxxxx>
Date: Thu, 13 Apr 2006 16:44:33 GMT

-----------------------------------------------------------------------------

---
Neil Truby wrote:
v480/StorEdge 3310 JBOD

My customer decided against having a cached RAID controller storage array on

the grounds of cost. Now they have quite acute performance problems at times,

wich the box running at 0% idel for up to 30 minutes. The database

So can anyone suggest how I might quantify the potential benefit of

substituting a RAID controller in the config?

This should be easy, a 'real' RAID controller is considerably less than more

software licenses or additional server hardware. Be sure to get a RAID

controller with a *battery backed* cache!

Obviously the database hardware requirements are dependent on the users

circumstances, but most often they are I/O bound not CPU bound....just as

your customer's sounds like.

Due to your customers money saving aspirations, they can now pay for a

re-installation of the database and OS on a properly designed storage system.

Storage, and the proper storage controllers, are *not* the first place to cut

costs - unless your customers data is worthless. Is it?
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: How long is a piece of string?
From: Neil Truby
References:
How long is a piece of string?
From: Neil Truby
Prev by Date: Re: lex %pointer not dynamical - workaround?
Next by Date: Re: fssnap gives error 22 on Solaris 8 system
Previous by thread: How long is a piece of string?
Next by thread: Re: How long is a piece of string?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: How long is a piece of string?
... Would you really install the OS on the new array though? ... RAID

controller
with a *battery backed* cache! ... My customer decided against having a

cached RAID controller
storage array ... Obviously the database hardware requirements are dependent

on the
users ... (comp.unix.solaris)
Re: 280R and A5200 question
... > You don't say what the purpose of this storage is, i.e. access stats.

...
array is going to be used to store scanned images that customers download ...

They, the
images, can then be downloaded by various ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
--------
Re: How long is a piece of string?

-----------------------------------------------------------------------------

---

From: "Neil Truby" <neil.truby@xxxxxxxxxxx>
Date: Thu, 13 Apr 2006 18:16:50 +0100

-----------------------------------------------------------------------------

---
Thanks for this.
Would you really install the OS on the new array though?
Surely it would be Ok to leave it on internal disk? Maybe move swap to the
RAID ..

This should be easy, a 'real' RAID controller is considerably less than

more software licenses or additional server hardware. Be sure to get a
RAID controller with a *battery backed* cache!

Right, but my dilema is not to persuade them to buy the RAID controller
instead of the cpus and licences, it's to persuade them to buy the RAID
controller rather than do nothing! Hence my need to try to qauntify the
likely benefit of doing so.

Cheers
--
Neil

"KJ" <here@xxxxxxxxx> wrote in message
news:Rdv%f.167431$g47.85879@xxxxxxxxxxxxxxxxxxxxxxxxxx

Neil Truby wrote:

v480/StorEdge 3310 JBOD

My customer decided against having a cached RAID controller storage array
on the grounds of cost. Now they have quite acute performance problems
at times, wich the box running at 0% idel for up to 30 minutes. The
database

So can anyone suggest how I might quantify the potential benefit of
substituting a RAID controller in the config?

This should be easy, a 'real' RAID controller is considerably less than
more software licenses or additional server hardware. Be sure to get a
RAID controller with a *battery backed* cache!

Obviously the database hardware requirements are dependent on the users
circumstances, but most often they are I/O bound not CPU bound....just as
your customer's sounds like.

Due to your customers money saving aspirations, they can now pay for a
re-installation of the database and OS on a properly designed storage
system.

Storage, and the proper storage controllers, are *not* the first place to
cut costs - unless your customers data is worthless. Is it?

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: How long is a piece of string?
From: Richard B. Gilbert
Re: How long is a piece of string?
From: Rick Jones
References:
How long is a piece of string?
From: Neil Truby
Re: How long is a piece of string?
From: KJ
Prev by Date: Re: fssnap gives error 22 on Solaris 8 system
Next by Date: Re: vold for solaris10
Previous by thread: Re: How long is a piece of string?
Next by thread: Re: How long is a piece of string?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: How long is a piece of string?
... My customer decided against having a cached RAID controller storage array

on the grounds
of cost. ... Due to your customers money saving aspirations, they can now pay

for
a re-installation of the database and OS on a properly designed storage

system. ... (comp.unix.solaris)
Re: First Mailbox Storage Error, need help with recovery
... but the PSS route will probably be the best way to go at this ... My

first thought
would be that there is something funky with the hardware. ... perhaps raid

controller
- dunno. ... I get ESENT: Catalog Database ...

(microsoft.public.exchange.admin)
Re: Adding second raid array
... I'll make an dedicated exchange only partition. ... >As Dave Nickason

said
a lot depends on your Raid controller. ... but make sure you do not mess up

your existing
array. ... >> you may find that all the drives operate at the speed of the
slowest. ... (microsoft.public.backoffice.smallbiz2000)
Re: ext3 filesystem not working with raid5
... and the array might already have reconstructed with bad data due ...

consider
*maxtor*250GIG IDE disks to be nice for single-user ... -rf' or a rooted

system or a RAID
controller running amuck at startup. ... > exactly what the problem is with

the ext3
filesystem and how to go ... (alt.linux)
RAID 1 XP Install issue
... onboard Promise 20265 Raid controller. ... I used to have a striped array
... I am now attemtping to use 2x160gb Maxtor drives (Is their ... routine

reports
no disks are present in my machine and the install fails. ...

(alt.comp.hardware.pc-homebuilt)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

----
Re: How long is a piece of string?

-----------------------------------------------------------------------------

---

From: Rick Jones <rick.jones2@xxxxxx>
Date: Thu, 13 Apr 2006 18:05:51 GMT

-----------------------------------------------------------------------------

---
A very apt subject. Anyone recognizing their question for what it is
like that deserves help :), so at the risk of helping a competitor,
doesn't Sun include CPU profiling tools with the OS? I would think
that with a little luck, you could take a profile of the kernel and
see how much CPU was being consumed by SW RAID routines. You then make
the broad handwaving ass-u-me-ption that a HW RAID controller would
make all of that go poof. And perhaps a little bit more since the
multiple I/Os to the JBODs for RAID writes would also go poof,
becoming "hidden" behind the HW RAID controller.

rick jones
--
oxymoron n, Hummer H2 with California Save Our Coasts and Oceans plates
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...
.

-----------------------------------------------------------------------------

---

References:
How long is a piece of string?
From: Neil Truby
Re: How long is a piece of string?
From: KJ
Re: How long is a piece of string?
From: Neil Truby
Prev by Date: Re: Are changes made with psradm persistent across reboot?
Next by Date: Re: Why use veritas?
Previous by thread: Re: How long is a piece of string?
Next by thread: Re: How long is a piece of string?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: My new Sun e250?
... Neil Rieck wrote: ... > Does Sun make/support a hardware based RAID
controller or is this the ... What did the application vendor say when you

asked them is there
was a VMS ... (comp.os.vms)
Re: Raid configuration for HP-UX
... vijay wrote: ... > 2> Raid controller for HP. ... Um, if
the Sun A1000 is really a RAID array, rather than just a JBOD ...

(comp.sys.hp.hpux)
Re: SCSI RAID Controller PCI for Sun SPARC
... > protected with RAID hardware solution. ... > I see that X6542A is Sun
Part Number for a SCSI RAID Controller SRC/P ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: How long is a piece of string?

-----------------------------------------------------------------------------

---

From: "Richard B. Gilbert" <rgilbert88@xxxxxxxxxxx>
Date: Thu, 13 Apr 2006 17:07:47 -0400

-----------------------------------------------------------------------------

---
Neil Truby wrote:

Thanks for this.
Would you really install the OS on the new array though?
Surely it would be Ok to leave it on internal disk? Maybe move swap to the

RAID ..

This should be easy, a 'real' RAID controller is considerably less than

more software licenses or additional server hardware. Be sure to get a RAID

controller with a *battery backed* cache!

Right, but my dilema is not to persuade them to buy the RAID controller

instead of the cpus and licences, it's to persuade them to buy the RAID

controller rather than do nothing! Hence my need to try to qauntify the

likely benefit of doing so.

Cheers

I have doubts about this. ISTR, that you said that CPU was maxed out. Unless

the system is devoting CPU time to Software RAID; e.g. SVM or Veritas,

hardware RAID is unlikely do do a thing for you! If you, in fact, relieve an

IO bottleneck with hardware RAID, you'll just be throwing more load on the

CPU and/or memory.

Additional CPUs are not guaranteed to help either. If the application is not

such that it can run in multiple processes, or it can't use threads,

additional CPUs won't be much help.

How long is a piece of string? Twice the distance from the middle to either

end!!

BTW, didn't the customer also by the cheapest machine they thought would do

the job? IMHO, when you buy a new machine it should have sufficient capacity

to handle at least three years of anticpated growth. When you outgrow it, it

becomes your test and development system and you buy a new box for

production.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: How long is a piece of string?
From: Neil Truby
References:
How long is a piece of string?
From: Neil Truby
Re: How long is a piece of string?
From: KJ
Re: How long is a piece of string?
From: Neil Truby
Prev by Date: Re: Solaris 9 DNS Config
Next by Date: Re: Solaris 9 DNS Config
Previous by thread: Re: How long is a piece of string?
Next by thread: Re: How long is a piece of string?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: How long is a piece of string?
... RAID controller with a *battery backed* cache! ... What you say of course

is
true as far as you go - the cpus may take up the slakc to their limit with

other types of activity
- but overall you'd expect the peaks to fewer and shorter. ... what do they

plan to do
in another six or eight months? ... (comp.unix.solaris)
Re: ATA Serielle 2 Controller
... Disable RAID controller in BIOS.. ... if the RAID BIOS setting
is ... (microsoft.public.windowsxp.hardware)
Re: Flashing lights on my disk drives on an Alphaserver ES40
... One thing to check for might be where are the disks plugged into? ... The

fact
that the lights are all 'flashing together' seems to indicate ... or whatever

the controller
is - even if it is not a RAID controller it ... (comp.os.vms)
Re: Flashing lights on my disk drives on an Alphaserver ES40
... One thing to check for might be where are the disks plugged into? ... The

fact
that the lights are all 'flashing together' seems to indicate ... or whatever

the controller
is - even if it is not a RAID controller it ... (comp.sys.dec)
Re: RAID 1 solution for desktop
... Usage will be almost entirely reading/writing large files, ... The

onboard
RAID on the MB(Asus P4C800 ... >adding a hardware RAID controller give

noticeable
performance benefits? ... >LSI MegaRAID, and 3Ware Escalade all seem

affordable. ...
(comp.arch.storage)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(04)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
-----
Re: How long is a piece of string?

-----------------------------------------------------------------------------

---

From: "Neil Truby" <neil.truby@xxxxxxxxxxx>
Date: Mon, 17 Apr 2006 13:01:33 +0100

-----------------------------------------------------------------------------

---
"Richard B. Gilbert" <rgilbert88@xxxxxxxxxxx> wrote in message
news:t-KdnXzk8aU-IKPZnZ2dnUVZ_vydnZ2d@xxxxxxxxxxxxxx

Neil Truby wrote:

Thanks for this.
Would you really install the OS on the new array though?
Surely it would be Ok to leave it on internal disk? Maybe move swap to
the
RAID ..

This should be easy, a 'real' RAID controller is considerably less than

more software licenses or additional server hardware. Be sure to get a
RAID controller with a *battery backed* cache!

Cheers

I have doubts about this. ISTR, that you said that CPU was maxed out.
Unless the system is devoting CPU time to Software RAID; e.g. SVM or
Veritas, hardware RAID is unlikely do do a thing for you! If you, in
fact, relieve an IO bottleneck with hardware RAID, you'll just be
throwing more load on the CPU and/or memory.

Well, the theory is that significant processing time IS being spent on
Solaris Volume Manager. But I'm trying to get a feel - from Sun, from this
peer group, anywhere really - on a typical amount.

I'm not sure I buy the second part of your argument. Funnily enough it's a
paraphrase of what the customer said: "I'd be disappointed if we don't see
at least 10% idle time at peak". What you say of course is true as far as
you go - the cpus may take up the slakc to their limit with other types of
activity - but overall you'd expect the peaks to fewer and shorter.

BTW, didn't the customer also by the cheapest machine they thought would
do the job? IMHO, when you buy a new machine it should have sufficient
capacity to handle at least three years of anticpated growth. When you
outgrow it, it becomes your test and development system and you buy a
new box for production.

They bought the box, then adequate, last summer but didn't anticipate the
growth!

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: How long is a piece of string?
From: Richard B. Gilbert
References:
How long is a piece of string?
From: Neil Truby
Re: How long is a piece of string?
From: KJ
Re: How long is a piece of string?
From: Neil Truby
Re: How long is a piece of string?
From: Richard B. Gilbert
Prev by Date: Re: EMC timefinder
Next by Date: Re: How long is a piece of string?
Previous by thread: Re: How long is a piece of string?
Next by thread: Re: How long is a piece of string?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Maximum price/performance
... the most CPU performance for my budget. ... If BIOS <7.0 the bios
has to be updated to v7.0. ... The RAID-0 should be set on NVIDIA RAID ...

The ASUS
web does not have the latest drivers for NVIDIA chipset ...

(alt.comp.hardware.pc-homebuilt)
Re: new PC build suggestions needed
... It runs realy verry fast. ... A lot of new technology comper to the

system
thar I bought 4 years ago: ... "pins" are on MB not on CPU ... Two Sata II HD
in raid o give very fast data transfer. ... (alt.comp.hardware.pc-homebuilt)
Re: raid advice?
... RAID-1 has very little CPU overhead, so there should be very little ...

difference
between hardware RAID and software RAID. ... The question isn't "what

filesystem
is best suited to RAID?", ... (comp.os.linux.hardware)
Re: What do you use?
... >> RAID server, though. ... Swap: 4081M Total, 240K Used, 4081M
Free ... The CPU usage is lower on the ICP system (remember that it has a

slower CPU ...
I noticed a drastic reduction of CPU load on the 3ware system when we went

... (freebsd-questions)
Re: Best HD setup for High Definition Video encoding >>????
... Unless you have parallelised encoding software, ... > going to add an Amd
X2 3800 dual core cpu in as well. ... For that you want lots of space and

low-latency
drives, ... > as a raid as the secondary drives. ...

(comp.sys.ibm.pc.hardware.storage)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
-----

4、Re: vold for solaris10

-----------------------------------------------------------------------------

---

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 12 Apr 2006 14:49:59 GMT

-----------------------------------------------------------------------------

---
"Joe" <spaceyjoe2020@xxxxxxxxx> writes:

hi,

how do you get the vold running for solaris10 on sunfirev210?
I checked the vold on another server v480 that has solaris10 and vold
is running

i tried the /usr/bin/vold start & doesnt seem to start.

I keep having to mount the cdrom. It isnt the ideal when I use the gui
installer and it asks for cd2- but the time i mount the cdrom with
disk2 the gui interface stops ruinning.

svcadm enable smserver

/etc/init.d/volmgt start
(or in later releases:

svcadm enable volfs

Casper
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: vold for solaris10
From: Joe
Re: vold for solaris10
From: greek_philosophizer
References:
vold for solaris10
From: Joe
Prev by Date: solaris management console
Next by Date: Re: root disk backup + restore options
Previous by thread: vold for solaris10
Next by thread: Re: vold for solaris10
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Burning on Solaris with cdrtools-2.01
... >the latest USB patch should work with the Sony USB DVD-RW. ... Wed Dec 8
15:20:28 2004 info: vold: running ... And I can't mount it. ... sudo format
-e /dev/rdsk/c3t0d0s2 ... (comp.unix.solaris)
Cant read/mount DVDROM
... vold daemon is not running on it & vold wont start even if I ... I've

verified
on another Sun machine to be a good media) drive ... Can't open boot device

...
mount: /dev/dsk/c3t0d0s0 is already mounted, /cdrom is busy, ...

(SunManagers)
summary: problem mounting music cds on Sol 9
... I'm having problems mounting music CDs on Solaris 9. ... Vold just

doesn't
mount the disk. ... can anyone recommend a good CD ripper for Solaris Sparc?

...
(SunManagers)
Re: Tips for getting a USB DVD writer working?
... Wed Dec 8 15:20:28 2004 info: label_cdrom: init ... Wed Dec 8 15:20:28

2004
info: vold: running ... And I can't mount it. ... Vendor_info: 'SONY
' ... (comp.sys.sun.hardware)
Problems with vold
... Running Solaris 9 on a Sunblade 100, and gnome desktop with all latest

... When
I put a CD in my CDROM drive, the drive spins and stops, but does ... mount

the
CDROM, ... I have added a "-v" to the vold process and monitor the vold.log

...
(SunManagers)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----Re: vold for solaris10

-----------------------------------------------------------------------------

---

From: greek_philosophizer@xxxxxxxxxxx
Date: 19 Apr 2006 15:08:23 -0700

-----------------------------------------------------------------------------

---
All true but the output of svcs -a does not tell you which service is
the removable media service.

If SUN chose a self descriptive name that would not matter.

If all you have is "smserver" you would probably not guess removable
media. I would guess sado/maso porn server. ( Not that I am into that
).

I suppose I will just write a little shell script that generates a list
of all the services, does a svcs -l on each one and picks out the
english description. That will have the same net effect.

-----------------------------------------------------------------------------

---

References:
vold for solaris10
From: Joe
Re: vold for solaris10
From: Casper H . S . Dik
Re: vold for solaris10
From: Joe
Re: vold for solaris10
From: greek_philosophizer
Re: vold for solaris10
From: Joe
Re: vold for solaris10
From: Dave Uhring
Re: vold for solaris10
From: greek_philosophizer
Re: vold for solaris10
From: Dave Uhring
Prev by Date: Re: FT_Get_BDF_Property: referenced symbol not found
Next by Date: Re: testing for boot block not working
Previous by thread: Re: vold for solaris10
Next by thread: solaris management console
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

5、Re: My FW/USB Solaris 10 experience

-----------------------------------------------------------------------------

---

From: "gerryt@" <gerryt@xxxxxxxxx>
Date: 12 Apr 2006 13:14:39 -0700

-----------------------------------------------------------------------------

---
balson@xxxxxxxxxxx wrote:

Hi, I have read some/most of the USB/firewire posts and and based on
the experiences of others, decided to give it a shot. I needed alot
more storage and SCSI storage was kinda out of the question basewd on
the amount of storage needed.
The Solaris experience wasn't as clean as that of the MAC, but I am
glad to say that this stuff does work. The USB interface works better
than does the firewire. At least for me it does.

I think thats the case for everyone.

The good:
- It works. I increased the strorage capacity of my Ultra 60 for alot
cheaper than purchasing equivalent sized SCSI drives. Speed is not an
issue for what I plan on storing on the drives. The USB drives are for
long term backup of lots of data. Will eventually go with 500gb drives
once the prices on those drives come down a bit.
- Can now swap in and out drives as need be without rebooting.
The bad::
- Using the USB interface, I can't have the unit powered on when
Soalris is booting. When I do, a message appears that says soemthing to
the effect that it can't connect to the device sitting at the other end
of the USB interface. After Solaris comes up, I can then turn the unit
on and it will see the drives. Both of them.

That could be your "tower"s busted firmware? I built up a Promise tower
last year (non USB admittedly) - had to turn off all RAID on it and
set the powerup state to "JBOD" - then used smc to configure it.
Even then the "LUNS" were restricted to 400 GB
So 2 were made instead of 1. Been in production for about a year now.

My direct attached stand alone USB drive has to be turned on first or
I cant mount it as a regular disk..

vfstab entry:
/dev/dsk/c3t0d0s0 /dev/rdsk/c3t0d0s0 /ion_160 ufs
1 yes -

- I can't have vold running. When vold is running, I can't see the
drives at all.

Again it could be broken firmware. I have vold running no problem

- Can't use the firewire interface. When I do, I can only see 1 of the
drives. Driver problem?

Maybe. I doubt its worth the bother to get FW going if USB works OK.

- When using the firewire interface, when I reboot Solaris and the unit
is powered on, I get a kernel panic. So I'm sticking with USB for now.

Thats expected behaviour on Solaris 10. Only 1 device is supported I
believe and thats a camera.. : > Im surprised it doesnt panic ANY time
you turn it on. but then Im using a Blade 1000/2000.

Maybe the next release of Solaris will correct the problem? If I wanted
to use the firewire interface, then I have to power the unit on after
Soalris has been booted.

You can try Nevada (Solaris 11) drivers - they worked for me.

-----------------------------------------------------------------------------

---

References:
My FW/USB Solaris 10 experience
From: balson
Prev by Date: Re: Why use veritas?
Next by Date: Re: Jumpstart process can´t find tape device
Previous by thread: Re: My FW/USB Solaris 10 experience
Next by thread: Help with this command
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Largest External HDD I Can Use With a Dimension 8200?
... Got the problem with the external Maxtor OneTouch II drive resolved. ...

I totally
uninstalled and reinstalled the Maxtor software, and used a different USB

port on the back of
my D-8300. ... These problems with ALL THE USB DEVICES started about 6 months

ago, and
increased with frequency until recently they became intolerable. ... I use

the two front USB ports
to connect my PNY Flash Drives, and get errors when plugging in to forward

connector, but not
the back one. ... (alt.sys.pc-clone.dell)
Re: Installing Fedora without a CDROM: my HOWTO
... Here is my HOWTO on installing Linux from a USB drive, ... disk or USB

key rather
than a CDROM boot disk. ... Put auxillary files (special drivers, post setup

install
config, ... This way I can mount any number of hard drives in the ...

(Fedora)
Re: usb hd
... >> however being a laptop i lack storage space. ... >> I
would like to use an external usb hard drive but am not sure of the ... have

a few unused
60 gig drives and hot swapper bays. ... (comp.os.linux.portable)
Re: Bootable USB HDD
... with the ability to boot from USB is required. ... both of whom stated
that USB external drives ... >>>>USB external hard drives. ...
(microsoft.public.windowsxp.hardware)
Re: Bootable USB HDD
... External USB Storage ... the boot order in the ... >that state
that "you can boot from a USB external hard ... both of whom stated that USB

external
drives ... (microsoft.public.windowsxp.hardware)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

6、Re: wait != -1 without child ?!

-----------------------------------------------------------------------------

---

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 11 Apr 2006 23:25:32 GMT

-----------------------------------------------------------------------------

---
Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

could someone please explain me, what is happening here?

thomas@azalin:~$ ./w.sh
p = 104425, r = 0
thomas@azalin:~$

Am I missing something obvious? I thought a.out doesn't fork, so it
should never be get something else than -1 from wait.

The shell constructs a pipeline such that the last process is the
parent of all other processes in the pipe.

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: wait != -1 without child ?!
From: Thomas Maier-Komor
Re: wait != -1 without child ?!
From: Thomas Maier-Komor
References:
wait != -1 without child ?!
From: Thomas Maier-Komor
Prev by Date: Re: V890 and hard disk
Next by Date: SunOs - Windows XP display redirect
Previous by thread: wait != -1 without child ?!
Next by thread: Re: wait != -1 without child ?!
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Do fork copies the "Code" Memory space?
... >and data are mapped to the same physical memory as in the parent. ...
Expressed in this posting are my opinions. ... to opinions held by my

employer, Sun
Microsystems. ... (comp.unix.programmer)
Re: Can anyone help? :Sparc 5 Solaris 1.1/SunOS 4.1.3
... Do you know what Rich Teer means by, ... able proper etiquette I will
not bother this newsgroup until I have honed my ... > to opinions held by my

employer,
Sun Microsystems. ... (comp.unix.solaris)
Re: High values for UID
... The internal disks are completely ... Pointless excercise if you want
to reduce space used. ... Expressed in this posting are my opinions. ... to

opinions
held by my employer, Sun Microsystems. ... (comp.unix.solaris)
Re: Any reason to use SPARC (vs. x86) these days?
... is not (supports serial console but without the ability to ... Expressed

in
this posting are my opinions. ... to opinions held by my employer, Sun

Microsystems.
... (comp.sys.sun.hardware)
Re: Can anyone help? :Sparc 5 Solaris 1.1/SunOS 4.1.3
... versions support SuperSPARC MP or any of the new graphics hardware ...

and select
the "SS5" (Under EOL systems) ... Expressed in this posting are my opinions.

...
to opinions held by my employer, Sun Microsystems. ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: wait != -1 without child ?!

-----------------------------------------------------------------------------

---

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 12 Apr 2006 08:52:38 GMT

-----------------------------------------------------------------------------

---
Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

Casper H.S. Dik wrote:

Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

could someone please explain me, what is happening here?

thomas@azalin:~$ ./w.sh
p = 104425, r = 0
thomas@azalin:~$

Am I missing something obvious? I thought a.out doesn't fork, so it
should never be get something else than -1 from wait.

The shell constructs a pipeline such that the last process is the
parent of all other processes in the pipe.

Casper

yes, that's OK. But the printed line is generated by the executable
a.out, which doesn't fork at all...

You don't understand; "a.out" may not fork() but the way the
pipe line is constructed the shell fork()s and then runs a.out
*in the parent*.

Like this:

if (fork() == 0) { /* Ok, start building pipe - child of shell */

pipe();
switch (fork()) {
case: 0 /* Child */
/* setup pipe */
/* run the first process in the pipeline */
exec(...)
_exit(1);
default: /* Parent */
/* setup pipe */
/* run the second process in the pipeline */
exec("a.out");
_exit(1);
case -1: perror("fork");
}
}

So the shell forks in the subprocess to build the pipeline; and then
uses that subprocess to execute the last process in the pipe (a.out).
That last process then has the other process as child.

sh -> a.out -> other process

and not what you might expect:

sh -> other process
-> a.out

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: wait != -1 without child ?!
From: Thomas Maier-Komor
References:
wait != -1 without child ?!
From: Thomas Maier-Komor
Re: wait != -1 without child ?!
From: Casper H . S . Dik
Re: wait != -1 without child ?!
From: Thomas Maier-Komor
Prev by Date: Re: wait != -1 without child ?!
Next by Date: Re: wait != -1 without child ?!
Previous by thread: Re: wait != -1 without child ?!
Next by thread: Re: wait != -1 without child ?!
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: C Program to execute programs in same console
... context of your parent shell. ... qdvd-author runs alls kinds of external
programs to generate ... process) current working directory, ...

(freebsd-questions)
Re: echo BOB | read name #not working?
... :>: subshell. ... a Bourne shell hidden away out of harm's reach.
... systems probably use a shell that executes the last element of a pipeline

...
(comp.unix.shell)
Re: Redirecting output from "kill" command?
... I realized it wasn't the kill command's output but the parent ... rather

than
the original interactive shell. ... the parent shell's stderr (my original

posting
where it said stdout was a typo) ... "Terminated" message for any jobs. ...

(comp.unix.misc)
Re: Fork, exec - setsid?
... >> Kill the shell that runs the pipeline. ... You have the pid.
... The whole thing has little to do with Perl, ... (comp.lang.perl.misc)
Re: Mike releases the "home" command, Version #2
... The CWD is always separate for every process. ... can change the parent

process
CWD. ... and which is already a shell builtin, ... > a troll, the
whole discussion is utterly moot. ... (comp.os.linux.misc)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: wait != -1 without child ?!

-----------------------------------------------------------------------------

---

From: Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx>
Date: Wed, 12 Apr 2006 11:06:25 +0200

-----------------------------------------------------------------------------

---
Casper H.S. Dik wrote:

Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

Casper H.S. Dik wrote:

Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

could someone please explain me, what is happening here?
thomas@azalin:~$ ./w.sh
p = 104425, r = 0
thomas@azalin:~$
Am I missing something obvious? I thought a.out doesn't fork, so it
should never be get something else than -1 from wait.

The shell constructs a pipeline such that the last process is the
parent of all other processes in the pipe.

Casper

yes, that's OK. But the printed line is generated by the executable
a.out, which doesn't fork at all...

You don't understand; "a.out" may not fork() but the way the
pipe line is constructed the shell fork()s and then runs a.out
*in the parent*.

I did understand what you mean. I just wanted to point out that this
kind of behavior is a little bit surprising, as a.out shouldn't need to
be aware of the context it is living in. But that seems to be an invalid
assumption on my part. As I pointed out in my other posts, /bin/sh seems
to be the only one surprising its children with this odd kind of
implementation. I agree, you save on fork. But I really care about the
unexpected gift of getting a pid from calling wait, when I wouldn't
expect it to return something else than -1.

BTW: I just x-checked /usr/xpg4/bin/sh: it returns -1. In contrast,
OSF1's /bin/sh behaves the same way as the one of Solaris.

Thanks anyway for your clarification. I guess I will have to go and
review some code concerning this issue...

Cheers,
Tom
.

-----------------------------------------------------------------------------

---

References:
wait != -1 without child ?!
From: Thomas Maier-Komor
Re: wait != -1 without child ?!
From: Casper H . S . Dik
Re: wait != -1 without child ?!
From: Thomas Maier-Komor
Re: wait != -1 without child ?!
From: Casper H . S . Dik
Prev by Date: Re: wait != -1 without child ?!
Next by Date: Printing Problem
Previous by thread: Re: wait != -1 without child ?!
Next by thread: Re: wait != -1 without child ?!
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: need function
... > I'd like to use pipe, fork and exec to capture stdout/stderr into a ...
fork and exec thing here ... > Texas Instruments ASIC Circuit Design

Methodlogy
Group ... (comp.unix.programmer)
Re: date | wc : fork, execve, dup, pipe problem!
... > File descriptors should be integers, ... Because i fork twice, this
code is just for testing, later i have a sitution ... > pipe() should be

called before
fork. ... instead of the syscall execve. ... (comp.unix.programmer)
Re: open3 and signals
... I came up with a solution using IO::Pipe and fork. ... I then talk to the

child
with the pipe, he runs the program, and talks ... parent process, which i

store
right after the fork before the loop. ... (comp.lang.perl.misc)
Re: Capturing another programs output into current C program...
... >My initial attempt was to use pipe and a call to fork to handle output

...
>redirection. ... This works fine for less than 10 iterations, ...

(comp.programming)
Re: page faults when spawning subprocesses
... Maybe using vfork rather than fork would help. ... parent to block until
the child has called execve, ... then you can do so manually using a pipe.

... (comp.os.linux.development.system)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(04)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
-------
Re: wait != -1 without child ?!

-----------------------------------------------------------------------------

---

From: Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx>
Date: Wed, 12 Apr 2006 10:39:29 +0200

-----------------------------------------------------------------------------

---
Casper H.S. Dik wrote:

Thomas Maier-Komor <thomas@xxxxxxxxxxxxxx> writes:

could someone please explain me, what is happening here?

thomas@azalin:~$ ./w.sh
p = 104425, r = 0
thomas@azalin:~$

Am I missing something obvious? I thought a.out doesn't fork, so it
should never be get something else than -1 from wait.

The shell constructs a pipeline such that the last process is the
parent of all other processes in the pipe.

Casper

Solaris' /bin/sh is different to ksh, bash, and zsh in this respect.
This is really confusing. Do I guess correctly that this is one of the
/bin/sh oddities that some people might rely on and that cannot be
changed because of backward compatibility guarantees?

Thanks,
Tom
.

-----------------------------------------------------------------------------

---

References:
wait != -1 without child ?!
From: Thomas Maier-Komor
Re: wait != -1 without child ?!
From: Casper H . S . Dik
Prev by Date: Re: wait != -1 without child ?!
Next by Date: Re: wait != -1 without child ?!
Previous by thread: Re: wait != -1 without child ?!
Next by thread: SunOs - Windows XP display redirect
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: wait != -1 without child ?!
... Casper H.S. Dik wrote: ... The shell constructs a pipeline
such that the last process is the ... (comp.unix.solaris)
Re: Sun Blade 1000 Opinions on Performance
... "Casper H.S. Dik" wrote: ... > As far as I know it is not a given
that these would be lies ... That notorious method is well known from the

politics.
... (comp.sys.sun.hardware)
Re: Linux Advocates Fear Solaris 10.
... Casper H. S. Dik wrote: ... > meant until it was hijacked by the open
source crowd. ... just like DEC's claim that renaming VMS to OpenVMS was

justified
... (comp.unix.solaris)
Re: strange behaviour of library in static initialization context
... Casper H.S. Dik wrote: ... MacOSX) or on LinuxThreads implementation.
... (comp.unix.solaris)
Re: source of boottime for uptime other than utmpx/wtmpx
... Casper H.S. Dik wrote: ... >>if you run ldd on that Kstat.so,
you will see that it links against ... >>I may have been a bit sloppy

implying that
you read a kernel variable. ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

随机文章：

0603月4，5，6日 2006-04-05

0507月28，29，30，31日 2005-09-08

0505月25，26，27日 2005-07-26

0505月7，8，9日 2005-07-03

0504月10，11，12日 2005-06-14

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0604月7，8，9日

mmmmn — Tue, 13 Jun 2006 22:49:09 +0800

1、Re: V890 and hard disk

-----------------------------------------------------------------------------

---

From: "Name goes here" <someone@xxxxxxxx>
Date: Tue, 11 Apr 2006 11:00:18 -0400

-----------------------------------------------------------------------------

---

"BL" <c50342182003@xxxxxxxxx> wrote in message
news:1144761910.64663***6190@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Hi,
I want to ask what command I should use in ok prompt to show all hard
disks when I boot up a sunfire V890? I try probe-scsi-all and nothing
dislay.

probe-scsi-all on V890's show all disk's FC included. You have to stop-A the
system after the self test and once the memory checks are done. V490's are
finicky this way too, and I guess all V class are the same. If you stop-A
them before, you may see nothing and the system is not bootable. You then
have to reset-all to get the system into a usable state. Another poster
claimed that devalias showed you all the devices, well thats not correct. It
only shows you device aliases. Thats not going to help you. But if you know
the h/w path to the device you think should be there by looking at the
show-devs output, you could walk the path and see whats there:

ok cd /pci@8,700000/SUNW,qlc@2,1
ok .properties
fc-boot-dev-portid 001412da
fc-boot-dev-portwwn 50 00 60 e8 02 9e cc 07
assigned-addresses 81001110 00000000 00000400 00000000 00000100
82001114 00000000 00102000 00000000
00002000
82001130 00000000 00140000 00000000
00020000
port-wwn 21 01 00 e0 8b 3c 8c ce
node-wwn 20 00 00 e0 8b 3c 8c ce
reg 00001100 00000000 00000000 00000000
00000000
01001110 00000000 00000000 00000000 00000100
02001114 00000000 00000000 00000000 00001000
compatible pci1077,2312.1077.10a.2
pci1077,2312.1077.10a
pci1077,10a

etc....

This will show you all sorts of wonderfull things, showing you that there is
a device is there. But probe-scsi-all is what you want. (It used to be
probe-fcal ).

Have at her.

-----------------------------------------------------------------------------

---

References:
V890 and hard disk
From: BL
Prev by Date: Re: V890 and hard disk
Next by Date: nfs access
Previous by thread: Re: V890 and hard disk
Next by thread: Binary package for Gnotime (Solaris 10)
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

2、Re: Managing Oracle workloads using SRM

-----------------------------------------------------------------------------

---

From: "Name goes here" <someone@xxxxxxxx>
Date: Tue, 11 Apr 2006 08:27:32 -0400

-----------------------------------------------------------------------------

---

"Eric Roland" <hbe@xxxxxxxxxxx> wrote in message
news:443AAB48.C69F3E95@xxxxxxxxxxxxxx

Hello,

Is it a simple way to manage workloads generated on solaris beetween
many oracle instances ?

Regards

E.R

Yes its simple enough once you wrap your head around how it all works. In
SRM you divide workloads into projects using the /etc/project file. Each
project can have any RCTL or resource control levels adjusted. You can even
create processor sets and assign a single or multi cpu processor set to a
workload or project using pools. See poolcfg. If you set the kernel
scheduler to FSS, you can divide up cpu workload's into shares for each
project (workload). You could name the project using the oracle instance
name. When you start the instance, you simply start it in that project using
the newtask command. SRM is a big area to cover is can do a lot of things in
many ways.

As an FYI, SRM in its former incantation was an add on module for Solaris 8.
Its a newer and better product and comes free with Solaris 9 & 10.

I have used it a lot, but not for Oracle but for search engines, zones
running small apps and web servers.

Good luck.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Managing Oracle workloads using SRM
From: Eric Roland
References:
Managing Oracle workloads using SRM
From: Eric Roland
Prev by Date: Help! Monitor connects to XVR-100 through HDF15 port no

display.
Next by Date: Re: How to make the second monitoron XBR-100 works?
Previous by thread: Re: Managing Oracle workloads using SRM
Next by thread: Re: Managing Oracle workloads using SRM
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(09)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
---
Re: Managing Oracle workloads using SRM

-----------------------------------------------------------------------------

---

From: Robert Milkowski <rmilkowski-NO-SPAM@xxxxxxxx>
Date: Fri, 14 Apr 2006 09:41:40 +0000 (UTC)

-----------------------------------------------------------------------------

---
Eric Roland <hbe@xxxxxxxxxxx> wrote:

The problem is especially how to use SRM beetweeb many oracle instances..

What problem?

Basicaly you have two options (in S10): put all Oracle instances in one
zone (could be glocal zone - so no need to creating any zones at all) and
then start each instance in a different project. Or put each instance
in its own local zone.

Look at http://www.sun.com/blueprints/browsesubject.html

--
Robert Milkowski
rmilkowskiASDASD@xxxxxxxx
http://milek.blogspot.com
.

-----------------------------------------------------------------------------

---

References:
Managing Oracle workloads using SRM
From: Eric Roland
Re: Managing Oracle workloads using SRM
From: Name goes here
Re: Managing Oracle workloads using SRM
From: Eric Roland
Prev by Date: Re: Monitor temperature?
Next by Date: Re: Why use veritas?
Previous by thread: Re: Managing Oracle workloads using SRM
Next by thread: Can't reboot after install Ultra 5
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

3、Re: TOP iowait vs iostat; 100% reading accurate?

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Mon, 10 Apr 2006 21:18:51 GMT

-----------------------------------------------------------------------------

---
Joe D. <newbie_from_newbie@xxxxxxxxx> wrote:

Good morning gentlemen;

I've seen several threads regarding I/O wait stats from the top
utility, and read some of the links that are referenced regarding how
I/O wait is calculated. Unless I've totally mis-read/mis-understood the
gist of the threads, I/O wait is not the "raise the flag; stop the
presses" issue that one might think it is at first glance, but rather
an indicator of CPU idle time on a healthy machine. I'm sure if I've
gotten this wrong, I will be corrected.

It's certainly not an immediate warning. It is an indicator of CPU idle
time. Whether or not the machine is healthy, iowait cannot tell you by
itself.

I would appreciate the collective's thoughts on the following scenario:

We have a V280R with 2 processors and 6 GB of RAM (recently upgraded
from 2 GB). This is a Bea Weblogic webserver front-ending a remote
Sybase DB, with no local storage other than the OS, but rather
everything is mounted via NFS from a Network Appliance filer.

We have a Compuware web-based utility (called VantageView) running on
the server keeping track of performance, with what looks like the top
utility. Ever since I upgraded the memory to 6 GB periodically, this
monitoring tool will show I/O wait at 100%. At the same time, the
system idle time will also be at 100%. So again, if I read the
aforementioned threads correctly, and I/O wait and system idle time are
inter-twined, then this kind of makes sense at first blush.

Do you have one tool showing 100% in two columns? Top shouldn't do
that. If a tool shows both iowait and idle, they should be disjoint
components of CPU idle.

Unfortunately, I cannot reproduce the same results running iostat; I
set a script to poll the system every 30 seconds with /bin/iostat
-czxn. I then parse the cpu stats from the output file, and I graph
them myself in excel. At no time do the wio numbers even remotely
approach the 100% mark as reported by this Compuware utility. The
normally running sar stats also do not show any real issues.

Perhaps the compuware utility is confused or calculating something its
own way. When I talk about the behavior of iowait, I'm referring only
to that calculated by the kernel and reported by 'top' and 'iostat'.
Your utility may mean something different.

The question I have is; has anyone else ever logged 100% WIO in top,
and seen such a discrepancy with the normal os based monitoring tools
such as iostat?

I haven't except with some older bugs. (I've seen 'top' read the wrong
values after a system patch changed some kernel variables).

Remember, iowait is a subset of idle CPU. If the CPU isn't idle, you
shouldn't havej iowait. Try running a cpu intensive job (like a stupid
perl or shell loop). If the iowait doesn't decrease below 100% by at
least the amount of user time, then something is wrong.

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

References:
TOP iowait vs iostat; 100% reading accurate?
From: Joe D.
Prev by Date: Re: TOP iowait vs iostat; 100% reading accurate?
Next by Date: Re: Managing Oracle workloads using SRM
Previous by thread: Re: TOP iowait vs iostat; 100% reading accurate?
Next by thread: SMC
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
IO WAIT Information From IBM
... the I/O wait metric in AIX. ... AIX scheduler, the CPU "queues", the
CPU states, and the idle or wait ... To summarize it in one sentence,

'iowait' is
the percentage ... (AIX-L)
Re: IO WAIT Information From IBM
... >understanding of how the I/O wait value is collected and calculated. ...
>of the CPU resource. ... The wait processes only job is to increment the

counters
that report ... >Each CPU can be in one of four states: user, sys, idle,

iowait. ...
(AIX-L)
Re: IO WAIT Information From IBM
... >understanding of how the I/O wait value is collected and calculated. ...
>of the CPU resource. ... The wait processes only job is to increment the

counters
that report ... >Each CPU can be in one of four states: user, sys, idle,

iowait. ...
(AIX-L)
Re: IO WAIT Information From IBM
... >understanding of how the I/O wait value is collected and calculated. ...
>of the CPU resource. ... >Each CPU can be in one of four states: user, sys,

idle,
iowait. ... >The tools print out the statistics using counters that the ...
(AIX-L)
Re: IO WAIT Information From IBM
... >understanding of how the I/O wait value is collected and calculated. ...
>of the CPU resource. ... >Each CPU can be in one of four states: user, sys,

idle,
iowait. ... >The tools print out the statistics using counters that the ...
(AIX-L)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

4、Re: VIPA on Solaris and HP

-----------------------------------------------------------------------------

---

From: Geoff Lane <zzassgl@xxxxxxxxxxxxxxxxxxx>
Date: 09 Apr 2006 11:44:13 GMT

-----------------------------------------------------------------------------

---
jbhamra@xxxxxxxxx wrote:

Hello Everybody,

What is an equivalent of AIX's VIPA (Virtual IP Address) on Solaris
and HP. Does anyone know how to configure it. Any docs. where I can get
the info will also be very helpful.

Not sure if it's the same but you can configure additional IPs to listen on
using ifconfig

# ifconfig nic:1 xxx.xxx.xxx.xxx netmask 255.255.255.0 up

or perhaps simpler for a temporary change

# ifconfig nic addif xxx.xxx.xxx.xxx netmask 255.255.255.0 up

If you want to make the change persist over a reboot create a file
/etc/hostname.nic:1 and put in a line containing the IP.

nic is the hardware interface name and depends on the hardware installed
which can be found from "ifconfig -a". More IPs can be configured by
incrementing the number after the colon.

--
Geoff Lane, Airstrip One

We spend billions on games of chance, not including weddings.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: VIPA on Solaris and HP
From: Sami Ketola
References:
VIPA on Solaris and HP
From: jbhamra
Prev by Date: VIPA on Solaris and HP
Next by Date: Re: Can Solaris 10 co-exist with Windows
Previous by thread: VIPA on Solaris and HP
Next by thread: Re: VIPA on Solaris and HP
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: VIPA on Solaris and HP

-----------------------------------------------------------------------------

---

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 10 Apr 2006 08:39:05 GMT

-----------------------------------------------------------------------------

---
Sami Ketola <Sami.Ketola@xxxxxxxxxxxxxxxxxxx> writes:

Geoff Lane <zzassgl@xxxxxxxxxxxxxxxxxxx> wrote:

jbhamra@xxxxxxxxx wrote:

Hello Everybody,

What is an equivalent of AIX's VIPA (Virtual IP Address) on Solaris
and HP. Does anyone know how to configure it. Any docs. where I can get
the info will also be very helpful.

Not sure if it's the same but you can configure additional IPs to listen on
using ifconfig

# ifconfig nic:1 xxx.xxx.xxx.xxx netmask 255.255.255.0 up

No. Not like this.

That actually does work as long as you "plumb" then first (assuming
we're talking about the Solaris, not the HP answer)

And this works for all releases of Solaris (not all requiring plumbing first)

or perhaps simpler for a temporary change

# ifconfig nic addif xxx.xxx.xxx.xxx netmask 255.255.255.0 up

This is the correct syntax. And it's not temporary

It's just as temporary as "ifconfig nic:1"

If you want to make the change persist over a reboot create a file
/etc/hostname.nic:1 and put in a line containing the IP.

No. The correct way would be to add second line to /etc/hostname.nic
containing the addif-directive.

Both will work. I'm not sure one or the other is "correct".

The nic:1 syntax fixes the interface name which is a useful property.

-----------------------------------------------------------------------------

---

References:
VIPA on Solaris and HP
From: jbhamra
Re: VIPA on Solaris and HP
From: Geoff Lane
Re: VIPA on Solaris and HP
From: Sami Ketola
Prev by Date: Re: Gnupg compile problems on Solaris 10/sparc
Next by Date: Re: Cold-reboot.
Previous by thread: Re: VIPA on Solaris and HP
Next by thread: Re: VIPA on Solaris and HP
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: md5
... some changes but it also needed to retain binary compatibility. ... it's

not
the default and not possible on Solaris 2.5.1 and before. ... Expressed in

this posting
are my opinions. ... to opinions held by my employer, Sun Microsystems. ...
(comp.unix.solaris)
Re: strange behaviour of library in static initialization context
... >Linux, FreeBSD, NetBSD and MacOSX, without any problems. ... not limited
to Solaris. ... Expressed in this posting are my opinions. ... to opinions
held by my employer, Sun Microsystems. ... (comp.unix.solaris)
Re: gethostans: buffer overflow exploit on solaris / sendmail
... >sendmail on solaris 9. ... >possible attempt to exploit buffer overflow
while looking up mail.xxx.be ... Expressed in this posting are my opinions.

...
to opinions held by my employer, Sun Microsystems. ... (comp.unix.solaris)
Re: Thread Local Storage
... >initialization is handled by the linker and the threading library, ... a
feature like this existed in the initial Solaris ... Expressed in this

posting are my opinions.
... to opinions held by my employer, Sun Microsystems. ...

(comp.unix.programmer)
Re: libncurses.so.5 open failure
... >I am new to Solaris 8 but I have some experience with Linux. ...
Expressed in this posting are my opinions. ... to opinions held by my

employer, Sun
Microsystems. ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(02)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
----
Re: VIPA on Solaris and HP

-----------------------------------------------------------------------------

---

From: James Carlson <james.d.carlson@xxxxxxx>
Date: 27 Apr 2006 08:09:51 -0400

-----------------------------------------------------------------------------

---
jbhamra@xxxxxxxxx writes:

What is an equivalent of AIX's VIPA (Virtual IP Address) on Solaris
and HP. Does anyone know how to configure it. Any docs. where I can get
the info will also be very helpful.

I don't know about HP, but Solaris 10 includes VIPA-like support via
Zebra routing. Check out /etc/sfw/zebra/README.Solaris, and look for
"HA SOLUTION" in that file.

Instead of using the lo0:1 interface, though, you should use vni and
the ifconfig usesrc keyword. See ifconfig(1M) for details.

--
James Carlson, KISS Network <james.d.carlson@xxxxxxx>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
.

-----------------------------------------------------------------------------

---

References:
VIPA on Solaris and HP
From: jbhamra
Prev by Date: Re: mpathd question
Next by Date: Re: ARP broadcast arrives, but MAC not changed
Previous by thread: Re: VIPA on Solaris and HP
Next by thread: USB and Swedish characters
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

5、Re: remote access from outside the domain

-----------------------------------------------------------------------------

---

From: Mike Delaney <mdelan@xxxxxxxxxxxx>
Date: Sat, 08 Apr 2006 01:52:07 -0500

-----------------------------------------------------------------------------

---
On 7 Apr 2006 17:28:05 -0700, lifesucks@xxxxxxxxx said something similar to:

I have a solaris 8 machine that i recently setup.

Out of curiousity, why did you do a new install of Solaris 8? The current
version is Solaris 10 and is avalible free of charge.

I am able to access it fine using telnet from with in the office from
all the machines and also access the webserver i set up on that
machine. But i'm not able to access it from outside the office network
. I cannot even ping the machine (IP addresses are public and other
win/macs are accessible from out side the n/w).

So it sounds like you have either no firewall in front of your network,
or at least an overly permissive one. This is an extremely bad idea, and
should be corrected immediately.

Since you can connect to your system from within the same subnet, can't
connect from outside the same subnet, and there's apparently no significant
firewall inbetween, the first thing I would check is your system's routing
table. Have you configured a default gateway? What does netstat -nr say?
What does netstat -nr say on one of the Windows or Mac clients that can
connect to outside systems?

I tried to find something on the net. I found information about editing
hosts.allow and hosts.deny which are supposed to be in the etc folder.
But I do not have them. Also there is some info on setting up the tcp
wrappers and then using these two access control files.

tcpwrappers is a 3rd-party software package designed to restrict access
to network services. Since it isn't shipped with Solaris 8, you would
have had to install it yourself in order for it to be part of your
problem.

Does that mean, by default telnet access will be denied from outside
the domain and you need to install tcp-wrappers (???) and then
configure these allow/deny files to be able to access from outside the
office?

You should never allow telnet access to a system from anywhere. It is
completely insecure. The same goes for rsh, rlogin, and ftp. For remote
shell access, install SSH. Both blastwave.org and sunfreeware.com provide
pre-built packages of OpenSSH.
.

-----------------------------------------------------------------------------

---

References:
remote access from outside the domain
From: lifesucks
Prev by Date: Solaris 10 jumpstart. Rules processed but profile isnt.
Next by Date: Re: Compiler for X64
Previous by thread: remote access from outside the domain
Next by thread: Solaris 10 jumpstart. Rules processed but profile isnt.
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Solaris 8 to Linux dump compatibility question (Nick Pettefar)
... Solaris 8 to Linux dump compatibility question ... FOLLOWUP: Netra X1:

How to
break the LOM prompt in order to ... patches not available to customers w/o

sunsolve
accounts... ... After this you will need to write a finish script that will

install the
same ... (SunManagers)
Solaris x86 FAQ 1/2
... Where can I obtain Solaris 2/x86 maintenance updates? ... What

information should
I have before an install? ... a 60GB disk shows up as only 28GB. ... Can
I create a partition for Solaris within my extended partition? ...

(comp.unix.solaris)
Re: trojaned SSHD ?
... I'd recommend moving to OpenSSH, which supports both ssh1 and ssh2 ...

platforms,
including Solaris. ... Information relevant to the installation of SSH on

NCMIR
systems. ... * Install Zlib 1.1.2 libraries, compiling from source, on

Solaris and IRIX
... (Focus-SUN)
Re: BSD Unix vs. Linux
... > and nss_ldap) necessary for network logins. ... quicker to install
than Windows for an experienced admin. ... Is FreeBSD that much more

difficult to install
and administer than ... Solaris on the desktop. ... (freebsd-newbies)
Cant upgrade Solaris 2.6 to Solaris 9 ? HELP
... I'm trying to upgrade one Sun Enterprise 3500 with Solaris 2.6 to new ...
install is in progress... ... Solaris Software Group: Entire Group ...

(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

6、Re: nfs link down, where do the files go?

-----------------------------------------------------------------------------

---

From: James Carlson <james.d.carlson@xxxxxxx>
Date: 07 Apr 2006 10:04:24 -0400

-----------------------------------------------------------------------------

---
getridofthespam@xxxxxxxxx writes:

I have a directory /myhost/logging that is nfs mounted
to a remote file system.
In that directory files are created. What happens while the
nfs link goes down. Could it happen that files will get created
on the local file system and become invisible forever when the
link comes up again? How should one handle this?

It's unclear what you mean by "nfs link goes down."

If you unmount that NFS file system, then you expose the mount point
directory. Anything written there goes into whatever the parent file
system is mounted -- possibly local disk. And if you mount over it
again (with NFS or any other file system), then those files will be
hidden away until you unmount again.

"Don't do that."

If the connection to the NFS server goes down, that doesn't cause the
mounted NFS file system to go away. And if it's down, then you just
can't write to anything in the NFS file system. So the problem
doesn't happen.

--
James Carlson, KISS Network <james.d.carlson@xxxxxxx>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
.

-----------------------------------------------------------------------------

---

References:
nfs link down, where do the files go?
From: getridofthespam
Prev by Date: Re: add RealPlayer to firefox
Next by Date: Re: bind port assignment
Previous by thread: nfs link down, where do the files go?
Next by thread: Re: nfs link down, where do the files go?
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
How can I determine version of apache running?? (pz4wfq .)
... NFS problem ... Determining who consumes disk space ... In a fact
I am writing this question here because everything is working fine with file

system shared on
Sun/Solaris box, while it is not working with file system shared on

EMC/Celerra. ... I am using
disk less clients that boots RedHat 8.0 over BOOTP and mount root file system

over nfs. ...
(SunManagers)
Re: Folder Synchronization
... Just to looking for AFS or Coda file system. ... and my server where
my documents are stored. ... > just to get NFS to work. ... > all that, but
when I am shutting down my laptop, one of the messages I ... (Fedora)
Re: Diskless Boot Problem
... > The root file system seems to mount correctly, but I am not sure how to

...
I don't have a memory file system set up, ... Check that you can mount the

nfs root
device. ... unconfigure the NIC before obtaining a new lease meaning that any

NFS ...
(freebsd-questions)
"Unpredictable" NFS errors?
... I use a simple backup script to tar major file systems onto a USB hard

drive ...
the 40 or so nodes to either FC2 or FC3), I encounter NFS statfs errors. ...

lose their
NFS file handles for one specific file system. ... Any insight you can

provide would
be greatly appreciated. ... (comp.os.linux.networking)
"Unpredictable" NFS errors?
... I use a simple backup script to tar major file systems onto a USB hard

drive ...
the 40 or so nodes to either FC2 or FC3), I encounter NFS statfs errors. ...

lose their
NFS file handles for one specific file system. ... Any insight you can

provide would
be greatly appreciated. ... (linux.redhat)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
---
Re: nfs link down, where do the files go?

-----------------------------------------------------------------------------

---

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Fri, 07 Apr 2006 15:46:56 GMT

-----------------------------------------------------------------------------

---
getridofthespam@xxxxxxxxx wrote:

Hi all,

As long as you don't use the 'soft' mounting option, the write should
block until the NFS server is available again. It won't write on the
local filesystem.

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

-----------------------------------------------------------------------------

---

References:
nfs link down, where do the files go?
From: getridofthespam
Prev by Date: IDE devices
Next by Date: panic[cpu3]/thread=140a000: vfs_mountroot: cannot mount root
Previous by thread: Re: nfs link down, where do the files go?
Next by thread: add RealPlayer to firefox
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
nfs link down, where do the files go?
... to a remote file system. ... In that directory files are created. ...
nfs link goes down. ... on the local file system and become invisible forever

when
the ... (comp.unix.solaris)
Re: nfs link down, where do the files go?
... In that directory files are created. ... on the local file system and
become invisible forever when the ... It's unclear what you mean by "nfs link

goes down."
... (comp.unix.solaris)
Re: Solaris kernel broken, maxgroups=16??
... NFS and the local file system. ... gets quite complex when one tries to

get
Windows locking ... reasonable approximation of the behaviour of a Windows

file server
in ... (comp.unix.solaris)
Re: How many files can you have in a VMS directory without performance

problems?
... > just the simple sequential file nature of directory files? ... My
50,000 ft view is that the ODS file system has a number of limitations ...

(comp.os.vms)
Re: Antivir doesnt work on local 5.x file system, but works fine on mounted

...
... > doesn't work on local file system in Freebsd 5.X ... > What is the

difference
between / and mounted file system, even if they are from the same ... (I'm

running it on
local file systems only - I'd try to avoid running it ... across a network

for a full scan
to avoid the network load, ... (freebsd-questions)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(06)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

8、Re: starting mysql

-----------------------------------------------------------------------------

---

From: Shea Martin <sheam@xxxxxxxx>
Date: Sat, 08 Apr 2006 02:40:29 GMT

-----------------------------------------------------------------------------

---
Thommy M. wrote:
See /etc/sfw/mysql/README.solaris.mysql

_OR_ http://www.sun.com/bigadmin/content/submitted/mysql_smf_tip.html

I think SMF is the way to go...

I just tried putting it in SMF, but I get this error when I try to enable it:

[ Apr 7 22:40:49 Executing start method ("/lib/svc/method/svc-mysql start") ]
svc.startd could not set context for method: chdir: No such file or directory
[ Apr 7 22:40:49 Method "start" exited with status 96 ]

svcprop shows that start/working_directory is :default. I assume that this is

the user's home dir. In this case the user is 'mysql', and the home dir is

/var/mysql. the perms on /var/mysql are 700, so root can't cd there.

I used svccfg to " setprop start/working_directory='/var' ". That made

everything work correctly. I thought I would post this info for other people

who follow that guide, and have trouble.

This can be fixed in the manifest by adding a working_directory attribute to

the method_context element, like so:
type='method'
name='start'
exec='/lib/svc/method/svc-mysql start'
timeout_seconds='-1'>
working_directory='/var'

Hope this helps someone.

~S
.

-----------------------------------------------------------------------------

---

References:
starting mysql
From: Shea Martin
Re: starting mysql
From: Dave Uhring
Re: starting mysql
From: Thommy M.
Prev by Date: Re: Solaris 10 / Express / 11 without GRUB?
Next by Date: Re: IDE devices
Previous by thread: Re: starting mysql
Next by thread: how to change Sun CLuster node hostname
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(07)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

9、Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem

-----------------------------------------------------------------------------

---

From: "Jansen Robert" <rjansen@xxxxxxxxx>
Date: Fri, 07 Apr 2006 18:27:30 +0200

-----------------------------------------------------------------------------

---
On Thu, 06 Apr 2006 18:35:12 +0200, John D Groenveld <groenvel@xxxxxxxxxxx>

wrote:

In article <e11of3$k1p$2@xxxxxxxxxxxxxx>,
Robert Jansen <rjansen@xxxxxxxxx> wrote:

RT 3.4.5 with RTFM-2.2.0RC2.

RT 3.4.5 is working fine for me with Apache2/mod_perl2.

Built Apache2 like so:
$ env PATH=/usr/bin:/usr/ccs/bin:/opt/SUNWspro/bin \
CC=cc CFLAGS=-xO3 CPPFLAGS="-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" \
./configure --with-ssl=/usr/sfw --enable-ssl --enable-proxy \
--enable-rewrite --enable-modules=all --enable-mods-shared=all \
--prefix=/opt/rt3/apache2
And edited apr_rules.mk and config_vars.mk in /opt/rt3/apache2/build
to add -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 to EXTRA_CPPFLAGS
and edited /opt/rt3/apache2/bin/apr-config and added the same to
CPPFLAGS and CFLAGS.

Built Perl like so:
$ env PATH=/usr/bin:/usr/ccs/bin:/opt/SUNWspro/bin \
./Configure -Dprefix=/opt/rt3/perl-5.8.8 -Dcc=cc -Duseshrplib \
-Uusemymalc -Duselargefiles -Ubincompat5005 -Doptimize=-fast

John
groenveld@xxxxxxx

Thanks for the reply.

Big difference with my machine is that you are using:
a) the Sun compilers
b) Apache 2 and Mod_perl2

And this is all on a Solaris 10 machine ?

question remains where the problem lies,...
Anyway,... thanks for the tipand recipe ,.. I guess I'll try with the Sun

compilers

I did some more digging around and came up with the following:

http://guest:guest@xxxxxxxxxxx/rt3/index.html?q=37376

!!!!!!!! It's a PERL BUG !! which should be fixed in perl 5.9.3.

A perl 5.9.3 default settings compile (only 64 bit int support expl.
chosen) with gcc 3.3.2 will show the following options set:
Compile-time options: PERL_DONT_CREATE_GVSV PERL_MALLOC_WRAP
PERL_USE_SAFE_PUTENV USE_64_BIT_INT
USE_LARGE_FILES USE_PERLIO

So I installed 5.9.3,... but,... and surely some other another problems
popped up. (2 RT problems mentioned below. but several required perl
modules for rt show strange test results under perl 5.9.3)

1) Tree::Simple has the "use 5.6.0" on line 4, which breaks RT under perl
5.9.3 (disabling the line in the .pm works around the bug)

2) from time to time,.... (not always !!) the "RT at a glance" page breaks
with the following:

HTTP/1.1 200 OK Date: Fri, 07 Apr 2006 14:02:31 GMT Server: Apache/1.3.34
(Unix) PHP/5.1.2 mod_perl/1.29 Pragma: no-cache Cache-control: no-cache
Keep-Alive: timeout=15, max=97 Connection: Keep-Alive, Keep-Alive
Transfer-Encoding: chunked, chunked Content-Type: text/html; charset=UTF-8
System error

error: Error during compilation of
/usr/local/packages/rt3/share/html/Elements/CollectionAsTable/ParseFormat:
Variable "$s" is not available at (re_eval 363) line 1.
Variable "$x" is not available at (re_eval 363) line 1.

context:
1: %# BEGIN BPS TAGGED BLOCK {{{
2: %#
3: %# COPYRIGHT:
4: %#
5: %# This software is Copyright (c) 1996-2005 Best Practical Solutions, LLC
....
code stack: /usr/local/lib/perl5/site_perl/5.9.3/HTML/Mason/Interp.pm:454
/usr/local/lib/perl5/site_perl/5.9.3/HTML/Mason/Request.pm:1088
/usr/local/lib/perl5/site_perl/5.9.3/HTML/Mason/Request.pm:1002
/usr/local/lib/perl5/site_perl/5.9.3/HTML/Mason/Request.pm:1197
/usr/local/packages/rt3/share/html/Elements/TicketList:120
/usr/local/packages/rt3/share/html/Elements/MyTickets:49
/usr/local/packages/rt3/share/html/Elements/TitleBox:48
/usr/local/packages/rt3/share/html/Elements/MyTickets:58
/usr/local/packages/rt3/share/html/index.html:80
/usr/local/packages/rt3/share/html/autohandler:215
raw error

Th rt.log shows similar messages:

(this is an example) [Fri Apr 7 14:18:42 2006] [warning]: Use of
uninitialized value in string eq at
/usr/local/packages/rt3/lib/RT/User_Overlay.pm line 1100.
(/usr/local/packages/rt3/lib/RT.pm:287)

By clicking on "Home" to reload the page displays the page correctly 9 out
of 10 )

It seemes that somehow perl is "loosing" variables (be it through expiry
and or pointer misalignment,... or else,... I don't know,...)

Comment much appreciated !

--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
From: John D Groenveld
References:
Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
From: Robert Jansen
Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
From: John D Groenveld
Prev by Date: Re: [Sun Blade 2500] regular can not login in with CDE
Next by Date: Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
Previous by thread: Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
Next by thread: Re: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

10、Re: performance of write access checks via proc(4)

-----------------------------------------------------------------------------

---

From: dkcombs@xxxxxxxxx (David Combs)
Date: Thu, 6 Apr 2006 00:22:36 +0000 (UTC)

-----------------------------------------------------------------------------

---
In article <dv2vnf$77v$1@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Claus Aßmann

wrote:

Thomas Maier-Komor wrote:

Have you considered using lock_lint(1) of Sun Studio? Maybe

It was on my list of things to do.

NOTE(MUTEX_PROTECTS_DATA(...)) could help you narrow down your problem.

It's an interesting tool with some problems. I will try to annotate my
program a bit more; currently lock_lint generates to many "false
positives".

Thanks for the suggestion.

FWIW, the language I use, "MAINSail" (not interpretive, but
with some intrepretive hooks, etc, built in) has a feature
in the debugger, where you can say "break after eg the 1millionth
instruction executed" (or maybe it's "statement"?).

So, if an otherwise undebuggable problem is *repeatable* (that's
*essential* for using this feature), then set the break count
to something high, so that it (mysteriously) fails BEFORE
it hits that count --

then you do it again, at HALF the count, etc

bisecting your way until you hit it --

then, knowing that, you set a breakpoint somewhat
BEFORE that, and single-step until you find the problem.

Sounds like a lot of runs, but of course bisection
converges so fast that it's better than it might sound.

Whether this can be done generally, eg in C, etc, I don't know.

David

-----------------------------------------------------------------------------

---

Prev by Date: Re: Logging who has logged in
Next by Date: Re: One of my servers is snooping another system...why?
Previous by thread: Re: Logging who has logged in
Next by thread: Apache/PHP 5.1.2/mod_perl Solaris 10 Problem
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

11、Re: One of my servers is snooping another system...why?

-----------------------------------------------------------------------------

---

From: tonij67@xxxxxxxxxxx
Date: 7 Apr 2006 12:05:54 -0700

-----------------------------------------------------------------------------

---
Atro Tossavainen wrote:

tonij67@xxxxxxxxxxx writes:

I checked the time frame that as pointed out and I noticed a snoop
process was running as the user nobody between the two systems.

It sounds probable that you might have been owned. You need to stop
the system dead (not even a proper "shutdown" but really kill it by
pulling the plug), move the system disk(s) to another system so that
you can take a backup without having to worry about the corruption in
the owned system altering whatever you plan to do with it, take the
backup (so that you can analyse what got you), wipe the disk, do a
clean reinstall from OS media, and apply the latest patches to prevent
this from happening again before connecting the system to the public
network again. Running Tripwire or something from now on might be
a good idea, too.

--
Atro Tossavainen (Mr.) / The Institute of Biotechnology at
Systems Analyst, Techno-Amish & / the University of Helsinki, Finland,
+358-9-19158939 UNIX Dinosaur / employs me, but my opinions are my own.
< URL : http : / / www . helsinki . fi / %7E atossava / > NO FILE ATTACHMENTS

I ran chkroot and it didnt turn up anything, but either way this system
is history. It's in the process of being replaced so I speeded up the
process :)

Does have me a bit concerned from here on out, tho.

-----------------------------------------------------------------------------

---

References:
One of my servers is snooping another system...why?
From: tonij67
Re: One of my servers is snooping another system...why?
From: Atro Tossavainen
Prev by Date: Re: T200 : did not get, what i bought :-)
Next by Date: Re: Daylight Savings time on Solaris 2.6
Previous by thread: Re: One of my servers is snooping another system...why?
Next by thread: nfs question
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(09)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

12、Re: install from solaris 10 media to RAID-1

-----------------------------------------------------------------------------

---

From: Wes W <wes@xxxxxxxxxxxxxxx>
Date: Tue, 04 Apr 2006 19:35:03 GMT

-----------------------------------------------------------------------------

---
Jonathan H N Chin wrote:
Assume a machine with two disks that one wishes to use as RAID-1.

I have been told that if one uses the install CD or DVD to install
Solaris-10 that it is impossible to create the mirrors first and
install "simultaneously" to both disks. Apparently one must install
onto one disk and then create the mirrors later (which will require
several reboots).

I have never installed off media but I find this amazing.
I know that jumpstart can do so, I see it doing it all the time.

Does the GUI installer really make it impossible and not just hard?
Can one really not just drop into a command shell and type the
relevant meta* commands and then tell the GUI installer to install
to /dev/md/dsk/d* instead of /dev/dsk/c?t?d?s? ?

-jonathan

Depending on your needs, i.e. server or workstation, I might suggest you try

the Solaris Express builds (Nevada b33 for example) if you are intending to

setup a workstation with data redundancy.

Just two days ago I wiped my Solaris 10 workstation install to use Nevada

(Solaris Express) b33 so I could use ZFS. Very impressed with ZFS. While you

can't, yet, install onto ZFS pools, that feature is planned and being worked

on. On amazing feature, it only takes one command to RAID my data slices!

Very nice.

# zpool create home c2d0t0s7 c3d1t0s7

Bam! RAID0 on two slices. If you want a mirror (RAID1), just add the word

'mirror' in there - see online documentation or ask if this is even remotely

suitable for your needs.

Obviously you can use ZFS on entire disks as well, but with only two hard

drives and ZFS boot (aka ZFS root) still in the works, this is my way around

that for now.
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: install from solaris 10 media to RAID-1
From: Jonathan H N Chin
References:
install from solaris 10 media to RAID-1
From: Jonathan H N Chin
Prev by Date: Re: Sun CC compiler: problem with vector
Next by Date: Re: Sun CC compiler: problem with vector
Previous by thread: Re: install from solaris 10 media to RAID-1
Next by thread: Re: install from solaris 10 media to RAID-1
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Solaris 8 to Linux dump compatibility question (Nick Pettefar)
... Solaris 8 to Linux dump compatibility question ... FOLLOWUP: Netra X1:

How to
break the LOM prompt in order to ... patches not available to customers w/o

sunsolve
accounts... ... After this you will need to write a finish script that will

install the
same ... (SunManagers)
Re: install from solaris 10 media to RAID-1
... install "simultaneously" to both disks. ... onto one disk and then create
the mirrors later ... I don't have any experience with RAID on Solaris 10 or

older,
but to setup data redundancy with Solaris 5.11 using ZFS setting up RAID1 can

be as easy as literally
one command: ... (comp.unix.solaris)
Solaris x86 FAQ 1/2
... Where can I obtain Solaris 2/x86 maintenance updates? ... What

information should
I have before an install? ... a 60GB disk shows up as only 28GB. ... Can
I create a partition for Solaris within my extended partition? ...

(comp.unix.solaris)
Re: trojaned SSHD ?
... I'd recommend moving to OpenSSH, which supports both ssh1 and ssh2 ...

platforms,
including Solaris. ... Information relevant to the installation of SSH on

NCMIR
systems. ... * Install Zlib 1.1.2 libraries, compiling from source, on

Solaris and IRIX
... (Focus-SUN)
Cant upgrade Solaris 2.6 to Solaris 9 ? HELP
... I'm trying to upgrade one Sun Enterprise 3500 with Solaris 2.6 to new ...
install is in progress... ... Solaris Software Group: Entire Group ...

(comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04
-----
Re: install from solaris 10 media to RAID-1

-----------------------------------------------------------------------------

---

From: KJ <here@xxxxxxxxx>
Date: Thu, 06 Apr 2006 14:49:12 GMT

-----------------------------------------------------------------------------

---
Jonathan H N Chin wrote:
Wes W <wes@xxxxxxxxxxxxxxx> writes:
Jonathan H N Chin wrote:
I have been told that if one uses the install CD or DVD to install
Solaris-10 that it is impossible to create the mirrors first and
install "simultaneously" to both disks. Apparently one must install
onto one disk and then create the mirrors later [...]

Depending on your needs, i.e. server or workstation, I might suggest you try

the Solaris Express builds (Nevada b33 for example) if you are intending to

setup a workstation with data redundancy.

Server or workstation, I don't mind. I'm just curious to know
if it really is impossible with the standard install media.
I'll be continuing to use my custom jumpstart scripts regardless.

-jonathan

I don't have any experience with RAID on Solaris 10 or older, but to setup

data redundancy with Solaris 5.11 (Express or newer) using ZFS setting up

RAID1 can be as easy as literally one command:

# zpool create mirror home c#d#t#s# c#d#t#s#

Then mount your new slice! Of course, you could setup RAID 0 the same without

the "mirror" switch.

http://www.opensolaris.org/os/community/zfs/docs/man_zpool/
.

-----------------------------------------------------------------------------

---

Follow-Ups:
Re: install from solaris 10 media to RAID-1
From: Darren Dunham
References:
install from solaris 10 media to RAID-1
From: Jonathan H N Chin
Re: install from solaris 10 media to RAID-1
From: Wes W
Re: install from solaris 10 media to RAID-1
From: Jonathan H N Chin
Prev by Date: Re: nslookup on solaris 10
Next by Date: Re: nslookup on solaris 10
Previous by thread: Re: install from solaris 10 media to RAID-1
Next by thread: Re: install from solaris 10 media to RAID-1
Index(es):
Date
Thread

-----------------------------------------------------------------------------

---

Relevant Pages
Re: Strange problem with Sun Blade 100 CD drive
... I buttoned the machine up and started a Solaris ... > move my data over

from
the original hard drive at a later time). ... I rebooted when the first part

of the install
was ... > It won't read either the Solaris 9 disks that I burned onto CD-Rs

...
(comp.unix.solaris)
Re: Disk labels (c0t0d0 & c1t0d => c2t0d0 & c3t0d0) weirdness
... change the disk labels when I install Solaris. ... I've used Solaris
for many months simply installing on disks "2" & "3" without any problem but

some programs, iostat
for example, are not as easily useful not having c0d0 and c0d1 populated

first as those drives are displayed
first. ... drive has one dedicated partition to Solaris on the fist

partition. ...
(comp.unix.solaris)
Re: Another Newbie question about an Ultra 5 and Solaris 10
... the disks have been wiped clean. ... > install Solaris 10? ...
> freeware) I should install/use besides what comes with Solaris 10? ... >

Dwayne
Lewis ... (comp.sys.sun.hardware)
Re: install from solaris 10 media to RAID-1
... I have been told that if one uses the install CD or DVD to install ...

onto
one disk and then create the mirrors later (which will require ... Depending

on your needs,
i.e. server or workstation, I might suggest you try the Solaris Express

builds if you are intending
to setup a workstation with data redundancy. ... Just two days ago I wiped my

Solaris 10 workstation
install to use Nevada b33 so I could use ZFS. ... (comp.unix.solaris)
Re: Solaris 8 -> 9 Upgrade with DiskSuite HOWTO?
... > If / is mirrored, you could just umount the other metadevices, upgrade

...
> solaris and /var on one of the disks in the mirror, ... stripes and mirrors
for /export, recreated them, and all was well ... ... The SDB slices are only

6.9MB, but there
is one on each of the 6 disks ... (comp.unix.solaris)

-----------------------------------------------------------------------------

---

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups

> comp.unix.solaris > 2006-04

随机文章：

0510月1，2，3日 2005-11-09

0508月4，5，6日 2005-09-11

0507月28，29，30，31日 2005-09-08

0506月28，29，30日 2005-08-14

0503月17，18，19日 2005-06-02

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0604月1，2，3，4，5，6日

mmmmn — Mon, 12 Jun 2006 23:58:06 +0800

1、Re: Dell Poweredge 830 64-bit sata raid

--------------------------------------------------------------------------------

From: Ian Collins <ian-news@xxxxxxxxxxx>
Date: Tue, 04 Apr 2006 18:27:37 +1200

--------------------------------------------------------------------------------
natG wrote:

Hi folks.
What are my chances that Solaris 10 will install on the Dell Poweredge 830
with SATA raid? (GB Ethernet.)

Probably 0.

SATA RAID isn't supported in Solaris 10.

If you can disable the RAID in the BOIS, Solaris Express might work,
depending on the controller chip.

The machine comes with several choices of operating systems, that reside
on the hard drive (Windows server 2003, RHL 64 bit, etc.) and can become
active/installed via a credit card. My question is, if I boot Solaris
installer off a dvd image, do I leave that stuff there? Should I low-level
format the drives first (via rom utility)?

Well that depends if you ever want to use them. You will have to create
a Solaris fdisk partition, but you won't have to reformat. If the
drives are setup as one partition, then you will have to lose the
current images.

Also, are my chances better with opensolaris. Why?

Yes, SATA support.

--
Ian Collins.
.

--------------------------------------------------------------------------------

Follow-Ups:
Re: Dell Poweredge 830 64-bit sata raid
From: llothar
Re: Dell Poweredge 830 64-bit sata raid
From: natG
References:
Dell Poweredge 830 64-bit sata raid
From: natG
Prev by Date: Re: Root file system / 100 % full implications..
Next by Date: Re: Buggy /usr/ccs/bin/ld on Solaris 8
Previous by thread: Dell Poweredge 830 64-bit sata raid
Next by thread: Re: Dell Poweredge 830 64-bit sata raid
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: Solaris OS versions supported on the cool thread series
... The software to be installed is supported on Solaris 9 only and AFAICT ... Solaris
9 mu8 CDs wont boot on this architecture. ... Its always been my experience that tech
support will bail on you if you ... So can or will Sun supply a version of Solaris 9 that
you can install ... (comp.sys.sun.hardware)
Solaris OS versions supported on the cool thread series
... The software to be installed is supported on Solaris 9 only and AFAICT ... Solaris
9 mu8 CDs wont boot on this architecture. ... Its always been my experience that tech
support will bail on you if you ... So can or will Sun supply a version of Solaris 9 that
you can install ... (comp.sys.sun.hardware)
SUMMARY: Jumpstart 64 bit problem
... I didn't get many replies but thanks to those who did. ... This was on Solaris
8 by the way. ... I am trying to install a client with 64 bit via jumpstart. ...
# To support the Network Time Protocol ... (SunManagers)
Re: SATA raid
... Which Adaptec SATA RAID would that be? ... Solaris 10 seems to atleast
support ... (comp.unix.solaris)
Re: Solaris OS versions supported on the cool thread series
... The software to be installed is supported on Solaris 9 only and AFAICT ... Solaris
9 mu8 CDs wont boot on this architecture. ... Its always been my experience that tech
support will bail on you if you ... So can or will Sun supply a version of Solaris 9 that
you can install ... (comp.sys.sun.hardware)

--------------------------------------------------------------------------------

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(01)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04
-----
Re: Dell Poweredge 830 64-bit sata raid

--------------------------------------------------------------------------------

From: Ian Collins <ian-news@xxxxxxxxxxx>
Date: Wed, 05 Apr 2006 11:03:55 +1200

--------------------------------------------------------------------------------
llothar wrote:

Do you mean that there is no SATA RAID or not even SATA in Solaris 10 ?

Are you replying to me? Please quote some context for those of us who
don't use google.

The SATA framework will be in Update 2, along with support for a number
of SATA controllers.

It is currently available in Solaris Express.

I don't think that will include any SATA RAID drivers. There are
drivers available form card vendors.

ZFS (also in U2) will go a long way to removing the need for SATA RAID.

--
Ian Collins.
.

--------------------------------------------------------------------------------

Follow-Ups:
Re: Dell Poweredge 830 64-bit sata raid
From: Michael Laajanen
References:
Dell Poweredge 830 64-bit sata raid
From: natG
Re: Dell Poweredge 830 64-bit sata raid
From: Ian Collins
Re: Dell Poweredge 830 64-bit sata raid
From: llothar
Prev by Date: Re: Maybe its is just me but....
Next by Date: Re: dd
Previous by thread: Re: Dell Poweredge 830 64-bit sata raid
Next by thread: Re: Dell Poweredge 830 64-bit sata raid
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: Boot from ICH6R Intel Software SATA Raid?
... installing a SATA DVD drive. ... CD/DVD drives is problematic at this
time. ... That board also has an ITE8212 as well as a Sil 3114R sata raid ... The
Sil 3114R controller seems to work just fine. ... (Fedora)
Re: A8N-SLI Deluxe : using SATA_RAID connectors just for extra SATA drives (not RAID) ?
... >> Thanks a lot, Paul, I've DL what you suggest. ... > "The Sil3114
S-ATA Controller option controls the state of the ... The SATA option allows the
tied SATA ... > ports to act as stand alone ports, while the SATA RAID option ...
(alt.comp.periphs.mainboard.asus)
Re: A8N-SLI Deluxe : using SATA_RAID connectors just for extra SATA drives (not RAID) ?
... > Thanks a lot, Paul, I've DL what you suggest. ... "The Sil3114 S-ATA
Controller option controls the state of the ... The SATA option allows the tied SATA
... Controller option set to the SATA RAID setting." ... (alt.comp.periphs.mainboard.asus)
Re: SATA vs SCSI RAID 5?
... > I recently setup a SATA RAID1 box on a highpoint controller. ... > Can
anyone tell me about SATA RAID 5 experiences? ... I not sure about the newer SATA RAID
controllers, ... usage and server load. ... (freebsd-isp)
RE: SATA Raid 5 recommendation?
... I've been talking with a local system integrator about doing SATA Raid 5 ...
recommending the promise controllers for Linux. ... (RedHat)

--------------------------------------------------------------------------------

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(07)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04
---------

2、Re: What does :kill really do in a SMF manifest?

--------------------------------------------------------------------------------

From: Dave Miner <dave.miner@xxxxxxx>
Date: Mon, 03 Apr 2006 15:52:14 -0400

--------------------------------------------------------------------------------
Gary Mills wrote:
In <Bm4Xf.43613$_S7.42935@xxxxxxxxxxxxxxxxxxxxxxxxxx> Darren Dunham <ddunham@xxxxxxxxxxxxxxxx> writes:

Gary Mills <mills@xxxxxxxxxxxxxxx> wrote:
I'll have to go back to a script that signals only the parent
process, then.

I just noticed this in one of the OpenSolaris discussions.

That is, svc.startd tores the contract ID of the service in the
"restarter/contract" property. You can use that ID with ctstat,
pgrep, pkill, or sigsend(2) to query and manipulate all of the
processes in the contract.

Ah, very nice integration. I'm impressed.

That, and the fact that 'pkill' has a -o option to signal only the
oldest process should make that script into a one-liner.

In this case, `pgrep -x -P 1 $SERVER' does what I want.

One note of caution: be careful with pgrep's, because if you run them in the global zone without qualifying them to be only the global zone, you'll get any processes in non-global zones, as well.

Dave
.

--------------------------------------------------------------------------------

Follow-Ups:
Re: What does :kill really do in a SMF manifest?
From: Henry Townsend
Prev by Date: Re: T2000 query temperature
Next by Date: The email attachment is not working properly on Solaris 8
Previous by thread: Sun Blade 100 will not boot.
Next by thread: Re: What does :kill really do in a SMF manifest?
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(05)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04
---------

3、Re: rcp + tar and gzip

--------------------------------------------------------------------------------

From: Darren Dunham <ddunham@xxxxxxxxxxxxxxxx>
Date: Mon, 03 Apr 2006 20:22:41 GMT

--------------------------------------------------------------------------------
merrittr <merrittr@xxxxxxxxx> wrote:

I was hoping to backup parts of a solaris filesystem using rcp, and
filtering it through tat to
tar and gzip it any idea how I would do this

rcp -r / | tar czvf output.tar.gz ????

You can't do it that way. Tar reads a filesystem, not an input stream.
Likewise 'rcp -r' is going to write to a filesystem, not to a stream.
You'd have to run tar on the remote filesystem directly and copy the
output.

rsh "tar cf - / | gzip -c" > output.tar.gz

--
Darren Dunham ddunham@xxxxxxxx
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >
.

--------------------------------------------------------------------------------

References:
rcp + tar and gzip
From: merrittr
Prev by Date: Re: Sun CC compiler: function overloads
Next by Date: Re: Sun CC compiler: function overloads
Previous by thread: rcp + tar and gzip
Next by thread: Trouble installing Sol8 on an Ultra 1E workstation
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: Quickest way to bulk copy many files from one disk cluster to another
... Copy the whole filesystem with tar while in production, ... Star
also implements a -copy flag that allows to avoid pipe related performance ... When one filesystem
is NFS mounted, try to make the _source_ filesystem ... (comp.unix.solaris)
Re: using star to move 3 millions files .. going slowly .. painfully
... On a fragmented filesystem, ... not to use logging but to set the "fastfs"
flag. ... Sun's tar is much slower than star. ... (comp.unix.solaris)
ISO vs .tar
... I am experimenting with a new backup strategy based on a DVD-RAM device, ...
I opt not to make a zipped archive, because unzipping it off the DVD-RAM ... straight .tar.
... Which type of "filesystem", the .iso ... (comp.os.linux.misc)
Re: dumping a samba directory
... prog like tar do you have any recommendations on which type of archiver ...
you can give dump the ... but it actually bypasses the normal filesystem ...
Samsung printers and HOSTING packages ... (freebsd-questions)
Re: Big time difference between (unix) find and DIRECTORY : why?
... Note also that tar doesn't have a problem with symbolic ... link from one filesystem
to another. ... Tim Cross ... My real e-mail is ... (comp.lang.lisp)

--------------------------------------------------------------------------------

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04

3、Re: T2000 query temperature

--------------------------------------------------------------------------------

From: Jorgen Moquist <jorgen.moquist@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Tue, 04 Apr 2006 01:27:09 +0200

--------------------------------------------------------------------------------
Fredrik Lundholm wrote:
In article <6XCXf.186$r64.189@xxxxxxxxxxxxxxxxxxx>,
Jorgen Moquist <jorgen.moquist@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:

there is no /usr/platform/`uname i`/sbin/scadm ,strange ?

Yes, this is strange, I wonder if anyone tried a V210/V240/V440 scadm command, if it works on sun4v.

Regs
Fredrik

tried:
t2000: # /usr/platform/SUNW,Sun-Fire-V440/sbin/scadm show
scadm: The SC hardware could not be initialized.

guess alom is different on a t2000, VxWorks and all. me thinks.
( dont know, they may have been VxWorks all the time )

i have serial+alom to the t2000, but scadm is preffered.

One can possibly use crossed cable from an unused ipge3 to the net-mgr port.
this will atleast give access to sc> showenvironment.

/jugge
.

--------------------------------------------------------------------------------

Follow-Ups:
Re: T2000 query temperature
From: hans m43
References:
Re: T2000 query temperature
From: Jorgen Moquist
Re: T2000 query temperature
From: Fredrik Lundholm
Prev by Date: Re: Sun CC compiler: function overloads
Next by Date: Re: Sun Blade 100 will not boot.
Previous by thread: Re: T2000 query temperature
Next by thread: Re: T2000 query temperature
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: T2000 query temperature
... Yes, this is strange, I wonder if anyone tried a V210/V240/V440 scadm ...
if it works on sun4v. ... The SC hardware could not be initialized. ... (comp.unix.solaris)
Re: T2000 query temperature
... Yes, this is strange, I wonder if anyone tried a V210/V240/V440 scadm command,
if it works on sun4v. ... (comp.unix.solaris)

--------------------------------------------------------------------------------

We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04
---------
Re: T2000 query temperature

--------------------------------------------------------------------------------

From: "cypherpunks" <cypherpunks@xxxxxxxxxxxxxx>
Date: 4 Apr 2006 00:26:23 -0700

--------------------------------------------------------------------------------
try this....

#prtpicl -v -c temperature-sensor

hope it works! let me know

--------------------------------------------------------------------------------

Follow-Ups:
Re: T2000 query temperature
From: Martin Paul
Prev by Date: Re: Memory Statistics - Kernel Tuning for File Cache?
Next by Date: Re: Sun CC compiler: function overloads
Previous by thread: Re: T2000 query temperature
Next by thread: Re: T2000 query temperature
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups > comp.unix.solaris > 2006-04
--------

5、Re: Looking for an USB PCI Card for my Ultra 10

--------------------------------------------------------------------------------

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 31 Mar 2006 22:39:06 GMT

--------------------------------------------------------------------------------
Casper H.S. Dik <Casper.Dik@xxxxxxx> writes:

The downloadable driver from gembird's website compiles on
Solaris (11b36) with:

env CFLAGS='-I/usr/sfw/include -L/usr/sfw/lib -R/usr/sfw/lib' ./configure

And also worked after I typed:

# update_drv -a -i '"usb4b4,fd11"' ugen

before plugging in the power brick:

SiS PM Control for Linux 1.1

(C) 2004,2005 by Mondrian Nuessle. This program is free software.
sispm_ctl comes with ABSOLUTELY NO WARRANTY; for details
see the file INSTALL. This is free software, and you are welcome
to redistribute it under certain conditions; see the file INSTALL
for details.

Found GEMBIRD SiS-PM on bus /dev/usb/4b4.fd11/0.

Usage information for sispm_ctl:

sispm_ctl [-q][-v] -o 1..4|all - switch outlet(s) on
sispm_ctl [-q][-v] -f 1..4|all - switch outlet(s) off
sispm_ctl [-q][-v] -g 1..4|all - get status
sispm_ctl [-q][-v] -b - switch buzzer on or off(?)

Option -q disables output for following commands specified on the command line
Option -v printfs detailed version and licence information
sispm_ctl return 0 on success, 1 on error and 2 if no SIS-PM compatible device could be found

libusb was added in Solaris 10 and that is what you need for the tool
to compile.

And to make it work. (Just tested it and it it's a nice toy)

--------------------------------------------------------------------------------

Prev by Date: Re: Solaris 10 Terminal Types
Next by Date: nohup not generating/updating nohup.out from cron
Previous by thread: Re: Solaris 10 Terminal Types
Next by thread: nohup not generating/updating nohup.out from cron
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: Can anyone help? :Sparc 5 Solaris 1.1/SunOS 4.1.3
... Do you know what Rich Teer means by, ... able proper etiquette I will
not bother this newsgroup until I have honed my ... > to opinions held by my employer,
Sun Microsystems. ... (comp.unix.solaris)
Re: High values for UID
... The internal disks are completely ... Pointless excercise if you want
to reduce space used. ... Expressed in this posting are my opinions. ... to opinions
held by my employer, Sun Microsystems. ... (comp.unix.solaris)
Re: Any reason to use SPARC (vs. x86) these days?
... is not (supports serial console but without the ability to ... Expressed in
this posting are my opinions. ... to opinions held by my employer, Sun Microsystems.
... (comp.sys.sun.hardware)
Re: Can anyone help? :Sparc 5 Solaris 1.1/SunOS 4.1.3
... versions support SuperSPARC MP or any of the new graphics hardware ... and select
the "SS5" (Under EOL systems) ... Expressed in this posting are my opinions. ...
to opinions held by my employer, Sun Microsystems. ... (comp.unix.solaris)
Re: fopen() limit on solaris
... resource limits or resource constraints. ... Expressed in this posting
are my opinions. ... to opinions held by my employer, Sun Microsystems. ...
(comp.unix.solaris)

--------------------------------------------------------------------------------

6、Re: Memory Statistics - Kernel Tuning for File Cache?

--------------------------------------------------------------------------------

From: Brian Utterback <brian.utterback@xxxxxxxxxxxxxxxx>
Date: Mon, 03 Apr 2006 13:52:43 -0400

--------------------------------------------------------------------------------
Peter Tribble wrote:
In article <e0r79g$t8h$1@xxxxxxxxxxxxxxx>,
Marc La Senne <mlasenne@xxxxxxxxxxx> writes:

Sorry, i forgot to say that the 2 servers are running without swap disk
The swap is in a ramdisk using ramdiskadm

root# ramdiskadm -a swap4 4g /dev/ramdisk/swap4
root# swap -a /dev/ramdisk/swap4

(Checks date of posting for evidence of April 1st...)

Doesn't this just waste 4G of memory?

Well, not exactly wasted, since a page of memory could be swapped
to it. However, it is exceedingly inefficient, since the same page
would not have had to be swapped if the 4GB had not been used by
the ramdisk. Since there is no swap device except the ramdisk,
it does no good at all to have the swap on the ramdisk, it just
adds a layer of indirection when a page is needed that has swapped
to the ramdisk.

On the other hand, your own comment about the inefficiency of running
without swap at all is somewhat misleading. While the amount of swap
that is reserved is non-zero and thus reduces the amount of physical
pages available for the lifetime of the reservation when there is no
swap disk, if you do not run out of memory, then it doesn't matter.

More to the point, if you run out of swap space, things go bad quickly;
programs abort, who knows what all. If you do not run out of swap space,
everything works fine. This is true whether or not you are using disk.
If you run out of swap, then you add more. So, if you are not using a
disk for swap, and never actually run out of swap, then it makes no
difference that you do not have a swap disk.

Granted, there is a small window of memory usage that the reservation
will put you over the top, but some percentage of those reserved
pages will end up being allocated and would have been needed regardless,
and thus made no difference.

Those people that do not configure a swap disk will make that choice
because either there is no disk available or they want to make sure
they never swap. In the former case, attaching a disk they do not
need would not be helpful, and in the latter case, they are more
likely to want to increase the memory size.

So, while it is true that configuring a small disk will avoid running
out of memory in a small number of cases, it is more likely that
it will turn out to be unnecessary or undesirable.

--
blu

Quidquid latine dictum sit, altum sonatur.
----------------------------------------------------------------------
Brian Utterback - OP/N1 RPE, Sun Microsystems, Inc.
Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom
.

--------------------------------------------------------------------------------

References:
Re: Memory Statistics - Kernel Tuning for File Cache?
From: Peter Tribble
Re: Memory Statistics - Kernel Tuning for File Cache?
From: Marc La Senne
Re: Memory Statistics - Kernel Tuning for File Cache?
From: Peter Tribble
Prev by Date: Re: SunPro for old Solaris releases
Next by Date: Re: Problem with 64 bit gnu ld on sparc (sol8)
Previous by thread: Re: Memory Statistics - Kernel Tuning for File Cache?
Next by thread: Re: Solaris install without install tools
Index(es):
Date
Thread

--------------------------------------------------------------------------------

Relevant Pages
Re: mdconfig unable to allocate memory
... > one ensure that all the memory is being used for suitable file caching ...
> disk, really; I assumed that it would immediately start dumping ... > anything
I put in it, into swap, causing disk access. ... file system with 4GB of RAM without
a problem, ... (freebsd-current)
Re: Q: How much Swap space?
... > plenty of unused physical memory. ... > discarding more frequently used
disk cache. ... The only "bad" thing about not having used any swap, IMHO, is that
I have ... (alt.os.linux.suse)
Re: Problem with adding more swap !
... > slices on the one disk for FreeBSD instead of having another ...
> you figure swap size needs by memory size. ... (freebsd-questions)
Re: Problems reclaiming VM cache = XFree86 startup annoyance
... => on the same overloaded disk. ... of paging that are initiated the
moment I invoke "startx". ... I executed "startx" the swap volume became very active (90--96%
busy; ... "Inactive" memory seems to get dumped over to "Cache" when I run ... (freebsd-stable)
Re: Does top work any more on Solaris ???
... My understanding is that Solaris manages free memory and swap partitions
... > to disk when there is insufficient RAM. ... (comp.sys.sun.admin)

--------------------------------------------------------------------------------

随机文章：

0603月4，5，6日 2006-04-05

0511月16，17，18日 2006-01-04

0511月4，5，6日 2005-12-20

0509月28，29，30日 2005-11-09

0507月28，29，30，31日 2005-09-08

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车

0603月25，26，27，28，29，30，31日

mmmmn — Sat, 22 Apr 2006 23:20:18 +0800

1、Re: i/o is high

------------------------------------------------------------------------------

From: "Dexthor" <gmolakal@xxxxxxxxx>
Date: 29 Mar 2006 06:19:00 -0800

------------------------------------------------------------------------------

--
r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b
device
325.6 13 2.9 0.1 0 3 0 8.9 0 98
A00000039d0
151.6 12.2 1.2 0.1 0 2.7 0 16.7 0 98
A00000038d0
888.8 1.2 17.2 0 0 2.3 0 2.5 1 98
A0000038Ed0

So, 17.2 MB is being read in 888.8 requests, so your avg: I/O request
size = 19.8 KB.
Your %busy is so high because, you are maxing out the Disk Queue with
large I/O request broken down into multiple small requests.

Here's what I would do:
1. Review sd_max_throttle setting that is applicable to your Disk
Subsystem and set it. EMC recommends 20.
2. vol_maxio - set this to something reasonable and relevant to your
disk. Suppose you have a SAN and using 2xGbit HBAs, then set this to
16000 (~16MB).
3. If you have striped your volumes today, then your problem is
probably linked to your Stripe Unit size. A large sequential I/O
request is broken down into multiple small requests(of size stripe
unit) by your Volume Manager Disk driver. If that is the case, you have
to restripe your stripe set with decent stripe unit size. In a Gbit/SAN
environment it is better to have StripeUnit == vol_maxio.

And talk to your Disk Management group to see if there are any
hotspots.

-Dexthor.

------------------------------------------------------------------------------

References:
i/o is high
From: atif76
Prev by Date: Re: Looking For 10/100 Ethernet Card With Solaris 8 Drivers
Next by Date: Re: New ssh/sshd patches for Solaris 9
Previous by thread: i/o is high
Next by thread: state in single user mode
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(03)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups >

comp.unix.solaris > 2006-03

2、Re: solaris big crash

------------------------------------------------------------------------------

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 28 Mar 2006 18:08:23 GMT

------------------------------------------------------------------------------

--
globulon@xxxxxxxxx writes:

we are running under Solaris 10 x86. Our system suddenly crashed a few
days ago. Symptoms are quite bizarre.
We found first that no ssh remote connection was possible. After a
reboot it was possible again for small amount of time to connect.

You are running an unpatched system and you've been hit but the
DST bug from hell.

You will need to set the date/time correctly from Solaris and then
reboot.

If the system only runs Solaris, I suggest you also run the following
command before adjusting the date:

rtc -c -z GMT

Then apply the appropriate patches.

Casper
.

------------------------------------------------------------------------------

Follow-Ups:
Re: solaris big crash
From: Wes W
References:
solaris big crash
From: globulon
Prev by Date: Re: Solaris Filesystem and Power Failures
Next by Date: Re: Solaris Root Mirroring
Previous by thread: Re: solaris big crash
Next by thread: Re: solaris big crash
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(08)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups >

comp.unix.solaris > 2006-03
-------
Re: solaris big crash

------------------------------------------------------------------------------

From: Casper H.S. Dik <Casper.Dik@xxxxxxx>
Date: 29 Mar 2006 08:53:23 GMT

------------------------------------------------------------------------------

--
Wes W <wes@xxxxxxxxxxxxxxx> writes:

Casper H.S. Dik wrote:

globulon@xxxxxxxxx writes:

You are running an unpatched system and you've been hit but the
DST bug from hell.

You will need to set the date/time correctly from Solaris and then
reboot.

If the system only runs Solaris, I suggest you also run the following
command before adjusting the date:

rtc -c -z GMT

Then apply the appropriate patches.

Casper

In addition to Casper's excellent insight, you may wish to join some of
Sun's free email newsletters regarding Solaris. The bug Casper
identified has been warned about or otherwise acknowledged for several
weeks at least.

Years, really; I think we warned both previous DST changes.

(the bug is limited to x64 running 64 bit kernels changing over to
timezones with a negative offset from GMT; a patch for this issue was
released in June 05)

------------------------------------------------------------------------------

Follow-Ups:
Re: solaris big crash
From: globulon
References:
solaris big crash
From: globulon
Re: solaris big crash
From: Casper H . S . Dik
Re: solaris big crash
From: Wes W
Prev by Date: Re: observe packets on Unix domain socket
Next by Date: Re: Looking for an USB PCI Card for my Ultra 10
Previous by thread: Re: solaris big crash
Next by thread: Re: solaris big crash
Index(es):
Date
Thread
We are proud to have Web Hosting and Rack Housing from 9 Net Avenue

Deutschland.
(10)
Security UNIX Linux Coding Shop Directory Usenet

Mailing-ListsNewsgroupsAboutPrivacySearch unix.derkeiler.com > Newsgroups >

comp.unix.solaris > 2006-03

随机文章：

0601月1，2，3日 2006-02-10

0511月1，2，3日 2005-12-19

0508月25，26，27日 2005-10-06

0506月16，17，18日 2005-08-06

0505月16，17，18日 2005-07-18

收藏到：Del.icio.us

博客大巴，你的个人传媒早班车