From ahoesch at smartsoft.de  Fri Feb  1 13:10:33 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Fri, 1 Feb 2008 19:10:33 +0100
Subject: Summary: Performance Problem
In-Reply-To: <3816C9DE-D0C1-11DC-9EF8-000393CA0072@smartsoft.de>
Message-ID: <FB014D48-D0F0-11DC-9EF8-000393CA0072@smartsoft.de>

Dear managers,

thanks to

Casper.Dik at Sun.COM
francisco roque <frisco at blackant.net>
Tim Bradshaw <tfb at tfeb.org>
"Matthew Stier" <Matthew.Stier at us.fujitsu.com>
Markus Mayer <mymaillists at gmx.at>

who suggested to use the following tools to see what is going on.

vmstat -p 10
      memory           page          executable      anonymous      
filesystem
    swap  free  re  mf  fr  de  sr  epi  epo  epf  api  apo  apf  fpi  
fpo  fpf
  18772512 4806232 774 1905 24 0 0  828    0    0    0    0    0  171   
28   24
  18861096 4880280 4072 706 13 0 0   35    0    0    8    0    0  540   
14   13
...

iostat
    tty        md0           md1           md3           md10           
cpu
  tin tout kps tps serv  kps tps serv  kps tps serv  kps tps serv   us 
sy wt id
    0   73 1673 117   27    9   0   39   21   1   34  1221  80   32   12 
  3  0 85


mpstat
CPU minf mjf xcal  intr ithr  csw icsw migr smtx  srw syscl  usr sys  
wt idl
   0  105   1   90    74   17 1431   57  299   73    2  2613   12   3   
0  85
   2  100   0  189   918  395 1434   44  304  116    2  2534   15   3   
0  82
  16  103   1  103    81   21 1467   61  294   85    2  2728   11   3   
0  86
  18   99   3  112   146   83 1663   66  333   86    2  3040   10   4   
0  85

pmap -S 22659
...
-------- ------- -------
total Kb 3284520 3251312

An oracle tool (em) currently - I am still running a test with a 
process currently using 3.2 GB of RAM - gives me

Memory utilization:	72 %

"prstat -s size" make me thinking that only 6 GB or so of the 16 GB 
installed in the machine are used:

    PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP
  22659 ahoesch  3208M 3194M cpu18    0    4   1:39:39  25% SmartClient/1
   5950 oracle   1151M 1082M sleep   59    0   0:00:15 0.0% oracle/13
   6793 oracle   1137M 1073M sleep   59    0   0:00:00 0.0% oracle/1
  ...
   6822 oracle   1137M 1077M sleep   59    0   0:00:00 0.0% oracle/1
   5944 oracle   1137M 1073M sleep   59    0   0:00:03 0.0% oracle/1
  27988 nbraun  465M  315M sleep   59    0   0:51:02 0.0% soffice.bin/6
...

I was told that the memory used by oracle is shared so that it actually 
uses 1150 MB only once. So this would be

	3.2 GB 	for my test process
	1 GB 	for oracle
	5 GB 	estimated for all the other processes
	=====================
	9 GB
	
Well, that's actually about 72 %. System performance is fine right now. 
It seems I can't reproduce the problem with my test process. Today a 
user with a 3.4 GB process bogged down the system completely. I will 
need to wait for the next real occurance and then recheck all the 
values.

What numbers in the above stats should I look for first? Looking at 
"prstat -s size" did not get me anywhere. Looking at the output made me 
always think that there should be enough memory left. Is anything fine 
until the second column of "vmstat -p 10" output reaches zero? mpstat 
does not seem to help either, it shows 85% idle all the time. What 
values in iostat output should I have an eye on when the problem 
reoccurs?

Thanks a lot,

   Andreas

> a user process utilizing 2GB - 4GB of memory on a SF 490 (2 x SPARC IV,
> 16 GB) causes serious performance problems for all users. It looks like
> much disk i/o is taking place. The performance problem goes away 1-2
> minutes after killing the user process that consumed the 2GB - 4GB of
> RAM. Since we have 16GB in the machine this should have no effect on
> other users IMHO. No swapping should  be performed since there is more
> than enough physical memory to hold all process data in RAM.
>
> What could be the problem? I heard that Solaris does not like greater
> portion of RAM being occupied by a single process. Is that correct? 
> Why?
>
> Thanks a lot,
>
>    Andreas
> _______________________________________________
> sunmanagers mailing list
> sunmanagers at sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From grove at zeta.org.au  Sun Feb  3 21:38:10 2008
From: grove at zeta.org.au (grove at zeta.org.au)
Date: Mon, 4 Feb 2008 13:38:10 +1100 (EST)
Subject: SUMMARY: Help with unbootable T2000 
Message-ID: <Pine.GSO.4.64.0802041327500.26783@zathras.uws.edu.au>

Hi,
thanks to everyone who replied.  Mostly it was reassuring that 
I could actually do this.   My big fear was that the T2000 would lose
all its SAN config, Zones and Zvols but as it turned out it was 
simpler in that respect than the actual restore.

I had 2 failed attempts at getting this system up again before coming across
an old Summary from 2006, that solved it for me.

http://www.sunmanagers.org/pipermail/summaries/2006-March/007141.html

As I found out, gtar seemed to lose something.  CPIO worked great.

Also, the LVM on Solaris 10 is a little more complex than SDS.

I had changed the vfstab to remove the LVM metadevs and edited 
the /etc/system to remove the rootdev entry, but failed to clear out 
/etc/lvm, which meant that I had a weird corruption and panic after 
my first attempt.

Clearing LVM completely and using a CPIO restore fixed the system for 
me and it is back in production.   I still have to put LVM mirroring 
back but I will wait till tomorrow.

Having a netboot really helps......   ....so does RSC/ALOM/CMT/&c....

Not being able to use Legato recover in a netboot scenario is a 
hassle.  I had to restore to another boot server that had space,
boot net -sw and then mount my restore  via NFS to cpio back to my 
root disk.   I guess if someone can make Legato work in a netboot
they will become a gazillionaire.   The gtar/cpio took many, many hours to 
complete....

Thanks again - this list is consistently an arse-saver.....
(ass-saver for the yanks)  ;)


rachel

-- 
Rachel Polanskis                 Kingswood, Greater Western Sydney, Australia
grove at zeta.org.au                http://www.zeta.org.au/~grove/grove.html
 		The price of greatness is responsibility.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Mon Feb  4 11:40:40 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Mon, 4 Feb 2008 17:40:40 +0100
Subject: Summary: Mounting disk after booting from cdrom
In-Reply-To: <6609877.1202144643208.SLOX.WebMail.wwwrun@ossi70.ossi.net>
Message-ID: <EB89B348-D33F-11DC-9C50-000393CA0072@smartsoft.de>

Dear mangers,

thanks to all your responses:

When booting from cdrom with "boot cdrom -s" the SVM stuff is not 
loaded so

	mount /dev/md/rdsk/d0 /a

does not work. However,

	mount /dev/dsk/cXtXdXs0 /a

should work. The reason why it did not for me first was that I used 
c0t0d0s0. A call of format revealed that I should have used c1t0d0s0 
instead. I could not mount

	mount /dev/dsk/c1t0d0s0 /a

either since the filesystem was unclean. Unfortunately "fsck -F ufs  
/dev/dsk/c1t0d0s0" was not able to repair the filesystem. I get 
"unrepairable cyinder". I have given up on c1t0t0d0. Thankfully the 
really valuable data is in slice 6. I will therefore reinstall Solaris 
10 in c1t0d0s0 and then mount /dev/dsk/c1t0d0s6 in order to copy the 
data over the network to another machine.

Thanks for all your help!

Regards,

   Andreas



>>> I am playing with a Solaris machine to get some practice in rescuing
>>> died file systems. I booted the machine from disk into single user 
>>> mode
>>> and ran
>>>
>>> 	fsck -F ufs /dev/md/rdsk/d0
>>>
>>> The file system check went through. I now want to mount the disk read
>>> write to e.g. reset the root password or delete /etc/defaultdomain. I
>>> tried
>>>
>>> 	mount /dev/md/rdsk/d0 /a
>>>
>>> and got "No block device". I tried
>>>
>>> 	mount /dev/md/dsk/d0 /a
>>>
>>> and got "Mount point /a" does not exist". I tried
>>>
>>> 	mount /dev/md/dsk/d0 /tmp
>>>
>>> and got "/tmp is already used". I also tried "mkdir /a" but this 
>>> fails
>>> with "filesystem is read only". :-(
>>>
>>> What am I doing wrong?
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Mon Feb  4 12:25:52 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Mon, 4 Feb 2008 18:25:52 +0100
Subject: Summary: NFS server in Solaris zone
In-Reply-To: <b86c876f0802040922p45f94259v7c46c4f08cd56ce3@mail.gmail.com>
Message-ID: <3BE8DF29-D346-11DC-9C50-000393CA0072@smartsoft.de>

Hi all,

NFS server works in the global zone only!

"Pedro Espinoza" <raindoctor at gmail.com> suggested to share 
/zonepath/root/local_fs from the global zone instead. I will try to get 
along without nfs in this case.

Thanks,

   Andreas

>> is that possible at all. I followed the usual approach
>>
>>         pico /etc/dfs/dfstab
>>         svcadm disable network/nfs/server
>>         svcadm enable network/nfs/server
>>         svcs -l network/nfs/server
>>
>> and get
>>
>> -bash-3.00# svcs -l network/nfs/server
>> fmri         svc:/network/nfs/server:default
>> name         NFS server
>> enabled      false (temporary)
>> state        online
>> next_state   disabled
>> state_time   Mon Feb 04 18:10:25 2008
>> logfile      /var/svc/log/network-nfs-server:default.log
>> restarter    svc:/system/svc/restarter:default
>> contract_id  383
>> dependency   require_any/error svc:/milestone/network (online)
>> dependency   require_all/error svc:/network/nfs/nlockmgr (online)
>> dependency   optional_all/error svc:/network/nfs/mapid (online)
>> dependency   require_all/restart svc:/network/rpc/bind (online)
>> dependency   optional_all/none svc:/network/rpc/keyserv (disabled)
>> dependency   optional_all/none svc:/network/rpc/gss (online)
>> dependency   optional_all/none svc:/network/shares/group (multiple)
>> dependency   require_all/error svc:/system/filesystem/local (online)
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From jesse-carroll at usa.net  Mon Feb  4 16:40:52 2008
From: jesse-carroll at usa.net (JESSE CARROLL)
Date: Mon, 04 Feb 2008 16:40:52 -0500
Subject: Summary: newfs with "-m 0"
Message-ID: <687mBDVn14270S07.1202161252@cmsweb07.cms.usa.net>

I got three responses, two which said no to "-m 0" and one self declared
essentially anonymous moderator (in an explicitly unmoderated forum) who
seemed to think that newfs is not a Solaris issue or is covered in the FAQ
some where (it isn't). The two answers where short on technical details but
supported my view that minimum free of 0 is a bad idea.

JC



------ Original Message ------
Received: Mon, 04 Feb 2008 11:41:57 AM EST
From: "JESSE CARROLL" <jesse-carroll at usa.net>
To: <sunmanagers at sunmanagers.org>
Subject: newfs with "-m 0"

All,

A colleague and I have been having a discussion about using the newfs
b-mb
option (AKA make_ufs bo free). In the past (on at least HPUX) using b-m
0b could lead to rather interesting results if the file system really goes
to 100%. I was told (more years ago than I want to think about) that b-m
1b is the lowest you should use. Using b-m 0b does create a bit larger
file system. What is a best practice for this?

JC
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ksonmez at oytek.com.tr  Tue Feb  5 01:57:11 2008
From: ksonmez at oytek.com.tr (Koray Sonmez)
Date: Tue, 5 Feb 2008 08:57:11 +0200
Subject: SUMMARY: C compiler for Solaris 10
In-Reply-To: <64879e520802041334l4a652aes9712adb7f32ef8a2@mail.gmail.com>
References: <60721B67EAF0994EAFFB561767B700140256558E@nets13ha.ww300.siemens.net>
	<96E68311E829E147865DC064BE1346CF954F94@oytexc01.oytek.com.tr>
	<64879e520802041334l4a652aes9712adb7f32ef8a2@mail.gmail.com>
Message-ID: <96E68311E829E147865DC064BE1346CF954FA1@oytexc01.oytek.com.tr>

Thanks everyone who replied.

I have a couple of choices.

1. Use Sun's compilers free of charge with Sun Studio 12:

http://developers.sun.com/sunstudio/downloads/index.jsp

2. Get the gcc binary packages from a third party site like
Sunfreeware or Blastwave:

http://www.sunfreeware.com/
http://www.blastwave.org/

regards,
koray

Bu e-mail'in tum icerigi gonderenin kisisel tasarrufu ile olusturulmus ve
gonderilmistir. Oyak Teknoloji Bilisim ve Kart Hizmetleri A.S. bu e-mail
icerigi hakkinda sorumluluk kabul etmez.
This e-mail has been sent by the
sender under his/her own individual discretion. Oyak Teknoloji Bilisim ve Kart
Hizmetleri A.S. accepts no responsibility.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From tfb at tfeb.org  Tue Feb  5 15:47:07 2008
From: tfb at tfeb.org (Tim Bradshaw)
Date: Tue, 5 Feb 2008 20:47:07 +0000
Subject: Summary: ZFS fsck
Message-ID: <0B4CD10C-F086-47FF-9CFB-186F3B3F672B@tfeb.org>

Thanks to Casper Dik, William D. Hathaway and Ryan C Anderson.

The conslusion is that: zfs import on a broken pool should not cause a  
panic (so this is a bug, I am not sure whether it is a known one), and  
that there is no very good way of looking at a pool which has not been  
imported to detect what, if anything is wrong with it.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From wib at informatik.uni-kiel.de  Wed Feb  6 03:00:37 2008
From: wib at informatik.uni-kiel.de (Willi Burmeister)
Date: Wed, 6 Feb 2008 09:00:37 +0100
Subject: SUMMARY: 127111-06 is declared as incompatible by 120185-01
In-Reply-To: <20080116080740.GA10038@falbala.informatik.uni-kiel.de>
References: <20080116080740.GA10038@falbala.informatik.uni-kiel.de>
Message-ID: <20080206080037.GA3792@falbala.informatik.uni-kiel.de>

Hi Sunmanagers,

> yesterday I tried to install patch 127111-06 on my Ultra 45 and
> got these irritating error message:
> 
>   Patch 127111-06 is declared as incompatible by 120185-01, 
>   which has already been installed on the system.
> 
> 120185-01 is a StarOffice Patch without any registered incompatible
> patch:

This was triggered by Bug 6404527. It's a missing blank in the pkginfo
file of nearly all StarOffice packages.

Here a quick workaround:

# cd /var/sadm/pkg
# grep -l ": R" */pkginfo | xargs perl -pi'.bak' -e 's/: R/:  R/'


Willi
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From tchipman at gmail.com  Wed Feb  6 19:14:22 2008
From: tchipman at gmail.com (Tim Chipman)
Date: Wed, 6 Feb 2008 20:14:22 -0400
Subject: SUMMARY: SSH Xforwarding broken after patches rolled - "known
	issue" - but not for Solaris 8
Message-ID: <bea9280b0802061614q6e3e8a18w622af41924b8253c@mail.gmail.com>

Hi All,

a quick summary. "Problem Solved". 				

Many thanks to Dennis Clarke, Allan West, Christopher Barnard for
their responses.

So, the story so far ..

-seems to be a problem with "older" builds of openssh, ie, a new build
of OpenSSH (or a pre-rolled binary of a new version from a 3rd party source,
such as blastwave) - should not exhibit this problem. Exact old vs new
that do or
don't exhibit the problem -- I haven't characterized exactly, sorry.

-originally reported in 2005 as a problem on Solaris9 and 10, but now
appears to be an issue on recent/fresh-patched Solaris8 machines also.
I haven't found
any mention of this on Sunsolve; maybe because Solaris 8 is EOL, and that
no sun-supported OpenSSH was bundled in Solaris8 (?)

-the workaround discussed online (not rolling back a patch, but instead
forcing SSHD to run only in IPv4 mode) - does work properly when implemented
right, ie, do both of these steps:

(1) edit sshd_config and specify an IPv4 format IP address for the line,
to read either

ListenAddress XXX.XXX.XXX.XXX (real ip address of system)

or

ListenAddress 0.0.0.0

should do the trick, PLUS,

(2) ensure sshd is started with a parameter passed to it, "-4", which
forces IPv4 mode, ie,
not including any kind of IPv6 mode (which is the default I gather?).
Then stop/restart ssh daemon and test.

One of the better (non-sun) links online I found was a discussion
online at the URL,

http://www.samag.com/documents/s=9915/sam0512i/0512i.htm


>From the replies I have gotten, this was a "known problem" at no less
than one other site.

Hopefully this summary will help other folks in the future who might
have this issue.


Many thanks,

---Tim Chipman


--------original posting below------------

I have  a Sparc Solaris8 (e250) which was patched last month
(~Jan-14-08) with all public (non-paying-sunsolve-support-user)
available patches for SolarisSparc (using the tool, "pca - patch check
advance).  Since that time, xforwarding no longer works from this
system via SSH.  (ie, in the past I would connect, ssh -X user at machine
- and have a functional xforwarding-via-ssh pipe back to my linux
desktop)

I note that there is an error message on the e250 machine now each
time this happens,

Feb  5 11:14:40 SERVERNAME sshd[1799]: [ID 800047 auth.error] error:
Failed to allocate internet-domain X11 display socket.

and a google search with this term brings up a number of hits
indicating this is a "known issue" on Solaris9 and Solaris10 machines
which were patched with patch ID 118305 - dating back to sometime in
2005.  There is a sunsolve entry visible at the URL,

http://sunsolve.sun.com/search/document.do?assetkey=1-26-101834-1

which discusses workarounds (don't use x-forwarding) or possibly
backing out the offending patch.

Alas, the sunsolve (and other) folks who saw this problem in ~2005
were all complaining on Sol9 and Sol10 boxes, and indicated (in
theory) that Solaris8 was not impacted.

I've already tried one workaround, attempting to force SSHD to start
up in ipv4 only mode (passing a -4 option to SSHD via the init.d/dir
startup script, and also via tweak in the sshd_config file to specify
an ipv4 format "listen" ip address explicitly).  Alas this kludge
doesn't appear to work. [[HINDSIGHT COMMENT - try to avoid typos!]]

Even more fun, since my system doesn't report having patch 118305
present, I don't have the option of following the sun recommended fix
of backing out that patch and appliying a slightly more down-rev
version which doesn't have this problem.

If anyone has any thoughts on how to work around this issue, any
pointers are greatly appreciated.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From mcmorran at mdibl.org  Fri Feb  8 13:37:58 2008
From: mcmorran at mdibl.org (Roy McMorran)
Date: Fri, 08 Feb 2008 13:37:58 -0500
Subject: SUMMARY: iSCSI LUN identification
In-Reply-To: <47AC89EF.8030108@mdibl.org>
References: <47AC89EF.8030108@mdibl.org>
Message-ID: <47ACA186.9090409@mdibl.org>

Roy McMorran wrote:
> Does anyone know of a way to drill down through scsi_vhci to identify 
> the specific iSCSI LUN used?  eg.
>   

Found it myself... nevermind.

iscsiadm list target -S

(missed the "-S" bit)

Thanks anyway ;-)
-- 

Roy McMorran
Systems Administrator
MDI Biological Laboratory
mcmorran at mdibl.org
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From rmp.dmd1229 at gmail.com  Fri Feb  8 13:49:05 2008
From: rmp.dmd1229 at gmail.com (rmp dmd)
Date: Fri, 8 Feb 2008 13:49:05 -0500
Subject: SUMMARY: Reboot problem
Message-ID: <61d255bb0802081049w5cc3b999y87804318c277e60f@mail.gmail.com>

Thanks for the replies:

here's what helped bring back the system:

   1.

    Determine which state database replicas are unavailable.

   # *metadb -i*

   2.

   **# *metadb -d c1t1d0s7*Verify that the replicas have been deleted.

   3.

   # *metadb*

   4.

   Reboot the system.

   # *reboot*


Re establish SVM.
On Feb 8, 2008 1:08 PM, rmp dmd <rmp.dmd1229 at gmail.com> wrote:

> Hi,
>
> After rebooting our sunfire V240 I'm getting error below.
>
>
> What can I do to resolv?
>
>
> Thanks!
>
>
>
> Rebooting with command: boot
> Boot device: disk0:a  File and args:
> SunOS Release 5.9 Version Generic_118558-39 64-bit
> Copyright 1983-2003 Sun Microsystems, Inc.  All rights reserved.
> Use is subject to license terms.
> e_ddi_get_dev_info: Illegal major device number <-1>WARNING: md: d20:
> (Unavailable) needs maintenance
> WARNING: Error writing ufs log state
> WARNING: ufs log for /mnt changed state to Error
> WARNING: Please umount(1M) /mnt and run fsck(1M)
> WARNING: forceload of misc/md_trans failed
> WARNING: forceload of misc/md_raid failed
> WARNING: forceload of misc/md_hotspares failed
> WARNING: forceload of misc/md_sp failed
> Hardware watchdog enabled
> configuring IPv4 interfaces: bge0.
> Hostname: redbull
> metainit: redbull: stale databases
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From scottd at HanoverDirect.com  Tue Feb 12 08:35:55 2008
From: scottd at HanoverDirect.com (Deiter, Scott)
Date: Tue, 12 Feb 2008 08:35:55 -0500
Subject: SUMMARY: Clone server
In-Reply-To: <ABF001998D701B4BBA6C58B131C42DFD09404BCC@HDIMAIL.hanoverdirect.lcl>
References: <274F5EE2-D4AA-11DC-A22C-000393CA0072@smartsoft.de>
	<ABF001998D701B4BBA6C58B131C42DFD09404BCC@HDIMAIL.hanoverdirect.lcl>
Message-ID: <ABF001998D701B4BBA6C58B131C42DFD09404BD9@HDIMAIL.hanoverdirect.lcl>

Thanks for all the responses.

We where able to get this system running by:
Booting from cd
  mounting the desired root disk as /xxx
  deleting the device tree ( /dev and /devices directories  along with
the path_to_inst file
  running /usr/sbin/devfsadm -r /xxx -p /xxx/etc/path_to_inst

  boot with -ar





Scott Deiter
System Administrator
Hanover Direct, Inc.
Hanover, PA
Voice: 717-633-3298
Fax: 717-633-3101
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From kkkong at ee.cuhk.edu.hk  Thu Feb 14 02:08:32 2008
From: kkkong at ee.cuhk.edu.hk (Alan Kong)
Date: Thu, 14 Feb 2008 15:08:32 +0800
Subject: Summary: Restored System Disk could not boot
In-Reply-To: <47B17107.9020806@ee.cuhk.edu.hk>
References: <47B17107.9020806@ee.cuhk.edu.hk>
Message-ID: <47B3E8F0.1050404@ee.cuhk.edu.hk>

Thank you to:

Damir Delija <ddelija at public.carnet.hr>
hike <mh1272 at gmail.com>
grove at zeta.org.au
"JULIAN, JOHN C (ATTSI)" <jj2195 at att.com>
jseymour at linxnet.com (Jim Seymour)

And Julian, John hit the head of the nail. Since Sun 280R uses fibre drive with world wide nos, the disk record needs to be removed before the new drive could be recongnised.

Procedures for re-building the device tree:

boot network or cdrom
mount new disk on /a
rm /a/dev/cfg/* /a/dev/rdsk/* /a/dev/dsk/*
devfsadm -C -v -r /a -p /a/etc/path_to_inst -t /a/etc/devlink.tab
touch /a/reconfigure
reboot -- -r 

Thank you.

Regards
Alan



Alan Kong wrote:
> Dear All,
>
> The system disk on our Sun Enterprise 280 server crashed and the disk 
> was replaced. The new disk was restored from tape and "installboot" ran. 
> However when the disk was booted, it kept asking to fsck the root 
> partition. The "fsck" failed when performed. As I had a second disk on 
> the server which was a clone of the system disk(the clone was a few 
> months old), I booted from this second disk and fsck the root partition 
> of the new disk. Then the server was rebooted with the new disk. Same 
> problem happened. The disks on the server were FC disks and the server 
> was running Solaris 2.8.
>
> Any suggestion on getting the system to run on the new disk?
>
> Thank you in advance.
>
> Regards
> Alan
> _______________________________________________
> sunmanagers mailing list
> sunmanagers at sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From speedyourmind at yahoo.com  Thu Feb 14 22:29:55 2008
From: speedyourmind at yahoo.com (Kiran Sharma)
Date: Thu, 14 Feb 2008 19:29:55 -0800 (PST)
Subject: Summary: unable to start ssh on solaris 10
In-Reply-To: <528210.88417.qm@web65415.mail.ac4.yahoo.com>
Message-ID: <633437.61547.qm@web65410.mail.ac4.yahoo.com>

Thanks to following individuals,
   
  Collo,Joel, Mohammed Uddin, Andrew Yauhwa Wu, Christopher Barnard, Alan Pae, Srivastava, Sanjaya 
   
  Basically the problem with the interface. I had two interfaces defined at /etc/hostname.hme0 and /etc/hostname.hme1 and both had same hostname specified. 
   
  When I boot it network/physical was not running. it was disable. I rebooted couple times and it failed each time which automatically disable ssh. 
   
  I removed hostname.hme1 and booted and it worked fine after that.
   
  Thanks all
  cheers
  KS

   
   
  
Kiran Sharma <speedyourmind at yahoo.com> wrote:
  Managers,
I installed solaris 10 6/06 and did live upgrade to 10 08/07. ssh was not working on 06/06 and after the upgrade it is not workin on 8/07 too. Any suggestions please,


bash-3.00# svcs -d network/ssh
STATE STIME FMRI
online 22:51:57 svc:/network/loopback:default
online 22:52:32 svc:/system/cryptosvc:default
online 22:58:49 svc:/system/filesystem/local:default
online 22:59:25 svc:/system/filesystem/autofs:default
online 23:09:16 svc:/system/utmp:default
maintenance 22:52:07 svc:/network/physical:default
bash-3.00#


bash-3.00# svcs -xv
svc:/network/physical:default (physical network interfaces)
State: maintenance since Wed Feb 13 22:52:07 2008
Reason: Start method exited with $SMF_EXIT_ERR_CONFIG.
See: http://sun.com/msg/SMF-8000-KS
See: man -M /usr/share/man -s 1M ifconfig
See: /etc/svc/volatile/network-physical:default.log
Impact: 6 dependent services are not running:
svc:/milestone/network:default
svc:/network/nfs/nlockmgr:default
svc:/network/nfs/client:default
svc:/network/nfs/status:default
svc:/application/management/sma:default
svc:/network/ssh:default



bash-3.00# svcs -a | grep -i ssh
offline 22:52:01 svc:/network/ssh:default
bash-3.00# svcs network/ssh
STATE STIME FMRI
offline 22:52:01 svc:/network/ssh:default
bash-3.00# svcadm enable network/ssh
bash-3.00#


---------------------------------
Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try it now.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers


       
---------------------------------
Never miss a thing.   Make Yahoo your homepage.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Sat Feb 16 06:48:31 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Sat, 16 Feb 2008 12:48:31 +0100
Subject: Summary: Solaris Express network config
In-Reply-To: <20080216050309.GA17609@porsche.suryo.com>
Message-ID: <18B19516-DC85-11DC-8B93-000393CA0072@smartsoft.de>

Hi all,

thanks to

"Luc I. Suryo" <luc at suryo.com>
Mike Brodbelt <mike at last.fm>
Udo Grabowski <udo.grabowski at imk.fzk.de>
A Darren Dunham <ddunham at taos.com>

There is this new service svc:/network/physical:nwam that should be 
disabled if one wants t configure a static ipaddress.

	svcadm disable svc:/network/physical:nwam
	svcadm enable svc:/network/physical:default

This alone did not do the trick for me. I had to enable and disable 
this nwam service again a few times. Suddenly it started to work. :-)

Thanks,

  Andreas
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ashish.madhavan at gmail.com  Tue Feb 19 09:29:46 2008
From: ashish.madhavan at gmail.com (Ashish Madhavan)
Date: Tue, 19 Feb 2008 19:59:46 +0530
Subject: Summary: Sun Cluster, vx mode - "mode: enabled: cluster
  inactive"
Message-ID: <56e4df110802190629s72e91900ua7d99b065b6734ec@mail.gmail.com>

Hi,
 I had to create rac-framework-rg for this to work. So it was not an install
issue just that the resources have to be setup :)

Thanks

On Feb 19, 2008 4:45 PM, Ashish Madhavan <ashish.madhavan at gmail.com> wrote:

> Some updates:
>
> VxVM version = 5.0_MP1_RP4.4
> ORCLudlm =  64bit 3.3.4.8 reentrant, async libskgxn2.so
>
> Basically issue is udlm daemon "dlmmon" isnt starting.
>
> Also I see the below errors
>
> cacao: Error: Fail to start cacao agent. (instance default)
> Error: Fail to start cacao agent. (instance default)
>
> AND messages file on nodeA shows the below error
> [ID 988885 daemon.error] libpnm error: can't connect to PNMd on nodeB
>
> Thanks,
> Ashish
>
> On Feb 19, 2008 11:57 AM, Ashish Madhavan <ashish.madhavan at gmail.com>
> wrote:
>
> >
> > Hi,
> >  I have installed sun cluster 3.2 on solaris 9 (Solaris 9 9/05). I want
> > to make it an active-active setup with shared veritas DGs. This setup also
> > has vxvm 5 with rolling pack 4 and solaris has all the latest pathes updated
> > via "updatemanager". The shared storage comes from DMX800.
> >  In order to get VxVM in cluster mode I have installed licenses for CVM,
> > VCS and also ORCLudlm package.
> >  The sun cluster install has all the necessary framework packages. But
> > the VX mode refuses to come in cluster mode:
> >
> > #vxdctl -c mode
> > mode: enabled: cluster inactive
> >
> > I am at my wits end on how to resolve this issue :(
> >
> > Any help is appreciated.
> >
> > Regards,
> > Ashish
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From mike at last.fm  Tue Feb 19 11:44:46 2008
From: mike at last.fm (Mike Brodbelt)
Date: Tue, 19 Feb 2008 16:44:46 +0000
Subject: SUMMARY: ELF Class error while compiling
In-Reply-To: <47BADE88.1040503@last.fm>
References: <47BADE88.1040503@last.fm>
Message-ID: <47BB077E.1090806@last.fm>

Mike Brodbelt wrote:

> # echo $LDFLAGS
> -L/usr/sfw/lib/amd64 -L/usr/local/lib/amd64 -R/usr/sfw/lib/amd64
> -R/usr/local/lib/amd64

<snip>

> So, I'm confused - it's building with a runtime link path of
> /usr/sfw/lib, and I can't seem to override this. Any advice would be
> much appreciated.

So, the problem turned out to be that the test program was linking
against a locally built 64-bit readline library, and when I'd built
that, I forgot to set the linker RPATH properly. Thanks to Francisco
Rocque who suggested some stuff that helped me find the problem. Lesson
learned - for runtime errors, "ldd -v" is your friend.

Mike
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From dreyerja at uni-paderborn.de  Thu Feb 21 05:02:41 2008
From: dreyerja at uni-paderborn.de (Jan Dreyer)
Date: Thu, 21 Feb 2008 11:02:41 +0100
Subject: SUMMARY: strange UFS-behaviour
In-Reply-To: <47B2F435.50705@uni-paderborn.de>
References: <47B2F435.50705@uni-paderborn.de>
Message-ID: <47BD4C41.70204@uni-paderborn.de>

Hi.

It seems we lost our filesystem, probably by a bug in the ufs-code,
maybe it's a hardwareproblem (scsi-connectors?), I didn't found out. At
the moment I'm rolling back our backup, which is better than nothing.

Anyway thanks for all of your responses:
- Some of you supposed a rootkit. I didn't think that was the problem.
First the developer of a rootkit won't make the complete content of a
directory vanishing (except it's a special dir with his stuff, but then
he will hide the directory itself!), second we did not observe any
unsusual network traffic or open ports, third we are behind a strong
company-firewall.

- there were no "overlay" mounts either. I unmounted the filesystem and
was not able to mount it again (whatever mountpoint I choosed):
IO-Error ...

so I definitely had a corrupt filesystem. Logging and Raid5 didn't save
my data :-(
fsck was NOT able to repair the filesystem either but was quitting with
error number 39:
-- snipp --
# fsck /dev/md/dsk/d55
** /dev/md/rdsk/d55
** Last Mounted on /var/something
** Phase 1 - Check Blocks and Sizes
newshadowclient: cannot malloc client array
# echo $?
39
-- snapp --

I equipped an old D1000 with some 146G drives and made a zpool (which I
should have done much earlier). Now the restore goes to this pool.

Greetings
Jan Dreyer


Jan Dreyer wrote:
> Hi @ll,
> 
> I watched a strange behaviour on one of our production servers (Solaris 
> 10u4).
> I have a DiskSuite-RAID5 mounted on, say, /var/something.
> After moving /var/something/subdir1 into /var/something/subdir2 our 
> /var/something vanished. Sort of ...
> 
> --snipp--
> # cd /var/something
> # pwd
> /var/something
> # ls -al
> total 0
> # # strange: . and .. are missing
> # cd subdir2
> # ls -al
> total 1358
> drwxrwxr-x   5 root     root         512 Feb 13 11:43 .
> drwxrwxrwx   0 user     group        512 Feb 13 11:43 ..
> drwxrwxr-x   3 user     group        512 Feb 13 11:43 subdir1
> [...]
> # mount|grep /var/something
> /var/something on /dev/md/dsk/d55 
> read/write/setuid/devices/intr/largefiles/logging/xattr/onerror=panic/dev=1540037 
> on Thu Jan 24 10:32:03 2008
> # metastat d55
> d55: RAID
>      State: Okay
>      Interlace: 32 blocks
>      Size: 860019840 blocks (410 GB)
> Original device:
>      Size: 860024544 blocks (410 GB)
>          Device     Start Block  Dbase        State Reloc  Hot Spare
>          c0t2d0s0       6090        No         Okay   Yes
>          c0t3d0s0       6090        No         Okay   Yes
>          c1t6d0s0       6090        No         Okay   Yes
>          c1t7d0s0       6090        No         Okay   Yes
> 
> Device Relocation Information:
> Device   Reloc  Device ID
> c0t2d0   Yes    id1,ssd at w2000000c5028c25f
> c0t3d0   Yes    id1,ssd at w2000000c5028c3aa
> c1t6d0   Yes    id1,ssd at w2000000c5028bfbe
> c1t7d0   Yes    id1,ssd at w2000000c5028c067
> -- snapp --
> 
> So everything seems ok, except that I can't see the directory-listing.
> fsdb gives me this:
> -- snipp --
> /dev/md/dsk/d55 > :ls
> /:
> ./              subdir2/
> ../             lost+found/
> -- snapp --
> 
> Anybody here who can advice me what to do? I fear to execute fsck 
> because the data on the disk is extremly valuable and tough there exists 
> a backup it's > 24h old and thats too much ...
> 
> As we watched this not the first time (the other were on the system disk 
> with Sol9 and we re-installed ...), I assume it may be a ufs-bug?
> 
> Thanks in advance ...
> Jan Dreyer
> _______________________________________________
> sunmanagers mailing list
> sunmanagers at sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From peter at ashlyn.co.uk  Thu Feb 21 06:15:26 2008
From: peter at ashlyn.co.uk (Peter Stokes)
Date: Thu, 21 Feb 2008 11:15:26 +0000
Subject: Summary - Sun B1600 how to reset switch admin password?
Message-ID: <47BD5D4E.8030000@ashlyn.co.uk>

Hi All

One response (thanks Tim), but no solution from the list, but a few web 
posts asking the same on Sun and other message boards.

I did manage to find the solution via friend of friend route. I will 
gladly pass on the info, but as it could also render the switch unusable 
if you delet the wrong file, I would prefer if amyone wanting to do it, 
to contact me directly. I am guessing this is why Sun have not released 
this in the public domain due to it not beinmg foolproof (unlike the 
password reset for the controller).

All the best

Peter


Hi All

Hopefully someone can put me out of my misery.

I have a B1600 which has a password set for the swich unit (not the
system control unit). I cannot find any way to reset the password as the
one already set is lost. I have reset the system controller password
using the remove power supply technique, but this does not appear to
have any impact on the switch password.

Any help greatfully received.

Thanks

Peter
----------------------------------------------------------------------------------------
Peter Stokes
Tel: +44 (0)1636 627900
----------------------------------------------------------------------------------------
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Thu Feb 21 11:23:45 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Thu, 21 Feb 2008 17:23:45 +0100
Subject: Summary: Backup home dirs on DVD
In-Reply-To: <29784997-DFB6-11DC-83F6-000393CA0072@smartsoft.de>
Message-ID: <5F9D2288-E099-11DC-A232-000393CA0072@smartsoft.de>

Dear managers,

thanks to

Barry Callahan <barryc at rjlsystems.com>
Anthony D'Atri <aad at dreamsnake.net>
"Luc I\. Suryo" <luc at suryo.com>
"Crist Clark" <Crist.Clark at globalstar.com>
A Darren Dunham <ddunham at taos.com>
francisco roque <frisco at blackant.net>
Deborah Crocker <crock at bama.ua.edu>

for their kind replies:

>>> we would like to backup ou home dirs on a DVD once a month. Solaris
>>> comes with this cdrw tool that we are used to use as follows
>>>
>>> 	cdrw -i /tmp/image.iso
>>>
>>> I know thatone can use something like
>>>
>>> 	/opt/sfw/bin/mkisofs -iso-level 4 -o /tmp/image.iso /home
>>>
>>> to create an iso image. However, we had already had problems with 
>>> long
>>> filenames using this approach. Is anybody actually using DVDs for
>>> backup purposes? How should we use mkisofs to make sure all files -
>>> even those with very lng file names - are included? It's ok if the 
>>> DVD
>>> can only be read by Solaris. Any idea?

We were suggested to put tar archive on the disk but that was not we 
had in mind. The following turned out to be a viable solution for us:

	mkdir /Carrier
	rm /Carrier/*
	ln -s  /home/projects /Carrier
	ln -s  /home/ahoesch/Development /Carrier
	ln -s  /home/beate /Carrier
	ln -s  /opt/FrontBase/Backups /Carrier
	ln -s  /opt/Backup /Carrier
	...

	mkisofs -uid 0 -gid 0 -N -l -R -U -allow-multidot -follow-links -d -D 
-v -o /tmp/image.iso /Carrier
	 /etc/init.d/volmgt stop
	cdrw -i /tmp/image.iso
	 /etc/init.d/volmgt start

This works like a charm. Thanks a lot for you help!

  - Andreas
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Loris.Serena at pfpc.ie  Sat Feb 23 06:32:37 2008
From: Loris.Serena at pfpc.ie (Loris.Serena at pfpc.ie)
Date: Sat, 23 Feb 2008 11:32:37 +0000
Subject: SUMMARY: Issue re-enabling ftp on jassed Solaris 10 8/07
In-Reply-To: <20080223111623.GB5297@baea.com.au>
Message-ID: <OF06AC535D.0ED0ACA9-ON802573F8.003F4E1A-802573F8.003F6949@pnc.com>

ooops...
Yes, that was it Brett!

Thanks a mill, that was both quick and spot on!

Loris






Brett Lymn <blymn at baesystems.com.au> 
23/02/2008 11:16

To
Loris.Serena at pfpc.ie
cc
sunmanagers at sunmanagers.org
Subject
Re: Issue re-enabling ftp on jassed Solaris 10 8/07






On Sat, Feb 23, 2008 at 10:30:42AM +0000, Loris.Serena at pfpc.ie wrote:
>
> [hercules]/$
> [hercules]/$ grep ftp /etc/hosts.allow
> in.tftpd:       ALL
> [hercules]/$

shouldn't that be in.ftpd?  You have allowed access to the tftpd
there.

-- 
Brett Lymn
"Warning:
The information contained in this email and any attached files is
confidential to BAE Systems Australia. If you are not the intended
recipient, any use, disclosure or copying of this email or any
attachments is expressly prohibited.  If you have received this email
in error, please notify us immediately. VIRUS: Every care has been
taken to ensure this email and its attachments are virus free,
however, any loss or damage incurred in using this email is not the
sender's responsibility.  It is your responsibility to ensure virus
checks are completed before installing any data sent in this email to
your computer."
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From lalit.kapoor at aricent.com  Sat Feb 23 14:36:29 2008
From: lalit.kapoor at aricent.com (Lalit Kapoor)
Date: Sun, 24 Feb 2008 01:06:29 +0530
Subject: Summary: NIS password problem with Xsession
Message-ID: <OF6B5173C0.C87AB2A2-ON652573F6.002C8F5B-652573F8.006BB41B@aricent.com>

Thanks for your responses.

It worked with dtlogin restart.


Regards,

Lalit Kapoor
Senior Product Specialist
 
A R I C E N T
 
Plot 31, Sector 18, Electronic City,
Gurgaon 122015, Haryana, India
 
Main     +91 124 234 6666 Ext. 2467
Fax      +91 124 245 5100
VOIP    6666 600 7303
 

 

***********************  Aricent-Restricted   ***********************

***********************  Aricent-Restricted   ***********************
"DISCLAIMER: This message is proprietary to Aricent  and is intended solely for the use of 
the individual to whom it is addressed. It may contain privileged or confidential information and should not be 
circulated or used for any purpose other than for what it is intended. If you have received this message in error, 
please notify the originator immediately. If you are not the intended recipient, you are notified that you are strictly
prohibited from using, copying, altering, or disclosing the contents of this message. Aricent accepts no responsibility for 
loss or damage arising from the use of the information transmitted by this email including damage from virus."
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From mymaillists at gmx.at  Tue Feb 26 06:15:43 2008
From: mymaillists at gmx.at (Markus Mayer)
Date: Tue, 26 Feb 2008 12:15:43 +0100
Subject: SUMMARY: Sun Cluster scdpmd (Disk Path Monitoring) high system
	load and core dumps
In-Reply-To: <200802261202.34589.mymaillists@gmx.at>
References: <200802261202.34589.mymaillists@gmx.at>
Message-ID: <200802261215.43680.mymaillists@gmx.at>

Great list!!!!

With thanks to Martin Pre_laber who responded almost instantly - We're dealing
with Bug ID 6563949,
http://sunsolve.sun.com/search/document.do?assetkey=1-66-201254-1

Impact:
Systems running Solaris Cluster 3.2, with the patches listed in
the "Contributing Factors" section below installed, may become extremely low
on memory causing other processes running on the system to be affected. It
may also be possible that "scdpmd" will dump core when it can't allocate
memory.

Workaround:
To work around this issue, disable "reboot_on_path_failure" property by using
the following command:
    # clnode set -p reboot_on_path_failure=disabled phys-foo-1

After changing this property, the load has reduces to almost 0, and the core
dumps should also stop.

Thanks!!!!


On Tuesday 26 February 2008, Markus Mayer wrote:
> Hi all,
>
> Hoping someone might be able to help here with two problems I have with
> scdpmd on a fresh Sun Cluster 3.2 installation on Solaris 10, update 3,
> with all patches installed that were available up until 15 February.  The
> systems have been "JASS'ed".
>
> First problem is that I have noticed that the scdpmd disk path monitor is
> eating on a V445 with 4 CPU's typically 22-25% of cpu resources.
> Essentially a whole cpu is being used just for disk path monitoring.  Right
> now there are 30 disks assigned to the system from a Hitachi AMS500, four
> paths to each disk, multipathing managed by MPxIO.
>
> The second problem is that scdpmd is dumping core quite often, at least
> every couple of days.  It does get restarted by the cluster infrastructure
> however the fact that is dumps bothers me.
>
> Does anyone know if there is any way to reduce the load, particularly as in
> the future I will get a lot more disks from storage assigned to the
> cluster, which would likely mean that the whole system resources would be
> eaten just with disk path monitoring.  According to the scdpmd manual, the
> monitoring could be shut off, however I have no idea what side effects that
> might have on the system or the cluster.
>
> Also, does anyone know anything that might help in resolving the core dumps
> that are happening with scdpmd.
>
> Thanks and regards
> Markus
> _______________________________________________
> sunmanagers mailing list
> sunmanagers at sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From johnladd68 at googlemail.com  Tue Feb 26 09:16:13 2008
From: johnladd68 at googlemail.com (John Ladd)
Date: Tue, 26 Feb 2008 15:16:13 +0100
Subject: Summary - How to force destruction of disk suite mirror while
	resync is in progress
Message-ID: <fa957010802260616ibc80ebawcc5be714aed5f473@mail.gmail.com>

Hello,

Thanks to those who provided suggestions.  In the end I found a way of
solving the problem, albeit quite brutal : disconnecting a volume from the
host in question.  The volumes are connected to the machine from a SAN, so
it was easy to unmap one half of the mirror, and then use "metadetach -f"
and "metaclear" commands to remove the devices which were now in error
state.  I then re-mapped the volume to the machine.

Thanks,

John.

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

Original question:

Hello,

I wish to completely destroy a disk suite mirror that I have mistakenly
created.  A resync is currently in progress between the two halves of the
mirror, and as the volumes are quite large this will take some time.  Also,
as I am experimenting with various configurations I don't wish to wait for
the resync each time.

Is there a way to force metaclear and/or metadetach commands despite the
resync?  The -f option does not work (...resync in progress).

I thought about destroying the meta database but as there are other
metadevices involved that I want to keep this is not an option.  Besides,
I'm not certain of the impact of doing this.

The mirror in question is under control of a metaset, I'm not sure if that
is relevant:

# metastat -s test1 -p
test1/d100 -m test1/d101 test1/d102 1
test1/d101 1 1 /dev/dsk/c4t1d0s0
test1/d102 1 1 /dev/dsk/c4t1d0s0

Maybe there is a way to destroy the metaset, thereby destroying the mirror
underneath?

John
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Wed Feb 27 18:13:37 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Thu, 28 Feb 2008 00:13:37 +0100
Subject: Summary: Solaris 10 Windows integration
In-Reply-To: <D7177194-E55A-11DC-9338-000393CA0072@smartsoft.de>
Message-ID: <A06B59DE-E589-11DC-9338-000393CA0072@smartsoft.de>

Dear managers,

thanks to

"Martin Marshall" <martinm at allwest.net>
renny.koshy at rubixinfotech.com
francisco roque <frisco at blackant.net>
<Gerhard.Weick at hzd.hessen.de>
Matthew Taylor <matthew.taylor at montgomerycollege.edu>
"Monappallil, George" <george.monappallil at rpfl.com>
sunmanagers at danielfrank.net
vogelke+sunlist at pobox.com (Karl Vogel)

for their kind responses. It's working for me now. Please find attached 
my notes.

Regards,

  Andreas


Samba
Samba is already preinstalled on Solaris 10.

	pkginfo -l | grep samba

       NAME:  samba - A Windows SMB/CIFS fileserver for UNIX (client)
       DESC:  samba - A Windows SMB/CIFS fileserver for UNIX (client) 
3.0.21b
       NAME:  samba - A Windows SMB/CIFS fileserver for UNIX (Root)
       DESC:  samba - A Windows SMB/CIFS fileserver for UNIX (Root) 
3.0.21b
       NAME:  samba - A Windows SMB/CIFS fileserver for UNIX (Usr)
       DESC:  samba - A Windows SMB/CIFS fileserver for UNIX (Usr) 
3.0.21b

We save the old rc* scripts as follows. Who knows may be we need them 
later.

	ls -l /etc/rc?.d/???samba

-rwxr--r--   6 root     sys          324 Jul  3  2006 
/etc/rc0.d/K03samba
-rwxr--r--   6 root     sys          324 Jul  3  2006 
/etc/rc1.d/K03samba
-rwxr--r--   6 root     sys          324 Jul  3  2006 
/etc/rc2.d/K03samba
-rwxr--r--   6 root     sys          324 Jul  3  2006 
/etc/rc3.d/S90samba
-rwxr--r--   6 root     sys          324 Jul  3  2006 
/etc/rcS.d/K03samba

	tar cvpf - /etc/rc?.d/???samba | gzip -c > /var/tmp/etc-rc-samba.tar.gz

This archive can be extracted again as follows if we need these files 
later.

	gunzip < /var/tmp/etc-rc-samba.tar.gz | tar -tvpf -

We can delete the scripts now but let's wait with that until the 
services are up and running.

	rm /etc/rc?.d/???samba

We create the sbm.conf file and make sure it looks as indicated below:

	cd /etc/sfw
	mv smb.conf smb.conf.orig
	cp smb.conf-example smb.conf
	pico /etc/sfw/smb.conf

[global]
         server string = Samba Server
         interfaces = 10.0.0.202
         bind interfaces only = Yes
         log level = 1
         syslog = 0
         log file = /var/log/samba/%m.log
         max log size = 50

[homes]
         path = /home/%u

[printers]
         comment = All Printers
         path = /var/spool/samba
         printable = Yes
         browseable = No

We create /var/log/samba before starting smb.

	mkdir /var/log/samba
	tail -f /var/log/samba/smbd.log

We used the start script supplied by Sun.

	pico /etc/init.d/samba

#!/bin/sh
#
case "$1" in
start)
	/usr/sfw/sbin/smbd -D
	/usr/sfw/sbin/nmbd -D
	;;
stop)
	pkill smbd
	pkill nmbd
	;;
*)
	echo "Usage: $0 { start | stop }"
	exit 1
	;;
esac
exit 0

	chmod a+x /etc/init.d/samba

We create a manifest file for setting up the service:

	pico /var/svc/manifest/network/samba.xml

<?xml version='1.0'?>
<!DOCTYPE service_bundle SYSTEM 
'/usr/share/lib/xml/dtd/service_bundle.dtd.1'>
<service_bundle type='manifest' name='WISWsamba:samba'>
   <service name='site/samba' type='service' version='0'>
     <create_default_instance enabled="false"/>
     <single_instance/>
     <!--
         First of all, if the config file is not present,
         then we needn't bother with anything else.
     -->
     <dependency name="config-file" grouping="require_all" 
restart_on="none" type="path">
       <service_fmri value="file:///etc/sfw/smb.conf"/>
     </dependency>

     <!--
       If there's no network, then there's no point in running
     -->
     <dependency name='loopback' grouping='require_all' 
restart_on='error' type='service'>
       <service_fmri value='svc:/network/loopback:default'/>
     </dependency>
     <dependency name='physical' grouping='optional_all' 
restart_on='error' type='service'>
       <service_fmri value='svc:/network/physical:default'/>
     </dependency>

     <!--
       Since Samba may be providing a home directory service,
       it is as well that we ensure that the file-systems are
       all mounted before it is started. This is not essential
       but in general it is a good thing and doesn't really hurt.
     -->
     <dependency name='fs-local' grouping='require_all' 
restart_on='none' type='service'>
       <service_fmri value='svc:/system/filesystem/local'/>
     </dependency>

     <!--
       now we have the start stop and refresh methods
     -->
     <exec_method name='start' type='method' exec='/etc/init.d/samba 
start' timeout_seconds='60'>
       <method_context/>
     </exec_method>
     <exec_method name='stop' type='method' exec='/etc/init.d/samba 
stop' timeout_seconds='60'>
       <method_context/>
     </exec_method>

     <property_group name='samba' type='application'>
       <stability value='Evolving'/>
     </property_group>
     <property_group name='startd' type='framework'>
       <propval name='ignore_error' type='astring' value='core,signal'/>
     </property_group>

     <stability value='Evolving'/>
     <!--
       A description of the Service
     -->
     <template>
       <common_name>
         <loctext xml:lang='C'>Samba Server</loctext>
       </common_name>
       <documentation>
         <manpage title='samba' section='1M'/>
         <doc_link name='samba.org' uri='http://httpd.samba.org'/>
       </documentation>
     </template>
   </service>
</service_bundle>

We doublecheck the manifest for errors and then start the service.

	svccfg validate /var/svc/manifest/network/samba.xml

	svcadm -v disable samba
	svccfg delete samba

	svccfg import /var/svc/manifest/network/samba.xml

	svcadm -v enable samba
	svcs -a | grep samba
	svcs -vx  svc:/site/samba:default
	svcs -l  svc:/site/samba:default
	svcadm restart svc:/system/svc/restarter:default
	ps -Af | grep smbd
	ps -Af | grep nmbd

We also add a service for the swat administration tool.

	pico /etc/services

swat            901/tcp                      # Samba Web Administration 
Tool

	pico /etc/inetd.conf

We add the following to /etc/inetd.conf

swat   stream   tcp   nowait   root   /usr/sfw/sbin/swat   
swat>>inet-swat

and run

	inetconv -f -o /var/tmp

This gives us

	swat -> /var/tmp/swat-tcp.xml
	Importing swat-tcp.xml ...Done

	inetadm -e svc:/network/swat/tcp:default
	svcs -a | grep swat
	svcs -l  svc:/network/swat/tcp:default
	inetadm -l svc:/network/swat/tcp:default
	
We can now access the samba administration page via the following URL:

	http://merkur:901

On the Solaris machine we create the samba user accounts we would like 
to grant access to samba resources to.

	smbpasswd -a ahoesch
	...

These accounts should already exist on the Windows Server and in 
/etc/passwd on the Solaris machine. On the Solaris machine we can do

	/usr/sfw/bin/smbclient -L SF490

to see which recources we have access to. It asks for a password of the 
current user which must be an existing samba user (smbpasswd -a ...), 
so if we run this command as root we should have done "smbpasswd -a 
root" before.

We can now log into the windows machine, e.g. using the Sun Ray windows 
connector

	/opt/SUNWutssc/bin/utssc -l de-DE -g 1024x768 <ipaddress of windows 
machine>

and then click on

	Start - Arbeitsplatz - Netzwerkumgebung - Arbeitsgruppencomputer 
anzeigen - Microsoft Windows Netwerk - Workgroup - Sambe Server (SF490) 
- homes

On any unix client with samba installed we could use

	smbclient //SF490/ahoesch

to access resources.

Troubleshooting
In case of problems check

	tail -f /var/log/samba/smbd.log

or try to start smbd manually with

	/usr/sfw/sbin/smbd -i

The following settings solved a "bind failed on port 139 socket_addr = 
0.0.0.0.  Error = Address already in use" problem for me.

[global]
         interfaces = 10.0.0.202
         bind interfaces only = Yes
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From joe_fletcher at btconnect.com  Wed Feb 27 21:57:45 2008
From: joe_fletcher at btconnect.com (joe fletcher)
Date: Thu, 28 Feb 2008 02:57:45 -0000
Subject: Summary: High tcp retransmits after adding secondary NIC pair
References: <4AE659A2271399438527DEE888E5440C034F03EB@HEMV2CUKER.he.local>
Message-ID: <4AE659A2271399438527DEE888E5440C034F03EE@HEMV2CUKER.he.local>

Hi,


In the end it turned out to be a bad switch. We replaced the switch hardware
and stats are now clean.

Cheers

Joe

----------------------------------- originally
---------------------------------------------------------------------


Got a 6900, 48CPUs, Solaris 10 118833-36. A couple of weeks ago we added in a
second pair of NICs with a view to segregating NFS traffic on one subnet from
the rest of the prod network. The machine has 4 cassinis (ce), one pair on
150
subnet and one pair on 152. IPMP one both pairs is active/standby. Netmasks
are fffffe00. Card speeds are 1000Base. Under periods of heavy load we see
high retransmission/duplicate packet rates. Clients are mostly v490s with ce
cards running at 100Base. The issue only seems to have appeared since we
added
in the extra cards.

Can anyone suggest places to look for possible cures? I'm wary of tuning the
stack initially but it remains an option.

TIA

Joe
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Crist.Clark at globalstar.com  Fri Feb 29 18:00:10 2008
From: Crist.Clark at globalstar.com (Crist Clark)
Date: Fri, 29 Feb 2008 15:00:10 -0800
Subject: SUMMARY: Getting PID of a Sub-Shell
In-Reply-To: <47C6C6CF.8C45.0097.0@globalstar.com>
References: <47C6C6CF.8C45.0097.0@globalstar.com>
Message-ID: <47C81DF9.8C45.0097.0@globalstar.com>

I got a number of responses. Two pointed to this,

 
http://www.unix.com/shell-programming-scripting/27767-ksh-how-does-line-child-sub-process-get-its-own-pid.html


Which verifies that others have had the same problem, but
does not provide a good workaround.

Someone else noted that this type of behavior is documented
in the manpage for PPID,

     PPID  This variable is set by the shell to the decimal  pro-
           cess  ID  of  the process that invoked the shell. In a
           subshell, PPID will be set to the same value  as  that
           of  the parent of the current shell. For example, echo
           $PPID and (echo $PPID) would produce the same value.

Which lends weight to the idea that it is a feature, not a bug.

I got a note from self-appointed list-cop saying it was off topic.
This was actually urgent even if I did not state it in the message.
I wanted to get the script in production before COB today. Yes, it
turned out not to be Solaris specific, but I had initially done an
(incorrect) test on another OS that made it look like it had
different behavior. Sorry on that account.

OTOH, I got two posts where people wanted to find out if I had
an answer and to please post a summary.

One response just said, "$!", which is a workaround when you
are running asynchronously, but can be a lot of overhead depending
on how you want to do it. Similarly, pulling the info out using
ptree(1), is a little clunky and not reliable if your parent
process is no longer around (you'd presumably be parsing output
of "ptree $$" which doesn't work after $$ is dead).

Finally, the best response mentioned that "$$" is set at an exec(2),
but not at later fork(2)s. So combining that with PPID, the trick
I'm going to use is,

	echo "Parent: $$"
	(
		MYPID=$(ksh -c 'echo $PPID')
		echo "Subshell: $MYPID"
	)

That is, run a command that tells you its parent's PID, which is the
process that called it.

Thanks.

On 2/28/2008 at 2:36 PM, "Crist Clark" <Crist.Clark at globalstar.com>
wrote:
> I've got a ksh script where I'd like to learn the PID of
> a sub-shell within the sub-shell code.
> 
> If you have code like,
> 
> 	$ cat pidtest.sh
> 	/bin/echo "$$"
> 	( /bin/echo "$$" )
> 	$ ksh pidtest.sh
>       29586
>       29586
> 
> You'll see the same number even though the second echo runs
> in a sub-shell with a different PID. If I run the script
> in truss(1) I do see that it is indeed happening as I expect
> it would, but the "$$" does not get changed when the sub-shell
> gets fork(2)ed off.
> 
> Bug or feature?
> 
> Anyway, anyone know of a way to work around it? How do I get
> the PID of that sub-shell from code within the sub-shell?


BB<information contained in this e-mail message is confidential, intended
only for the use of the individual or entity named above. If the reader
of this e-mail is not the intended recipient, or the employee or agent
responsible to deliver it to the intended recipient, you are hereby
notified that any review, dissemination, distribution or copying of this
communication is strictly prohibited. If you have received this e-mail
in error, please contact postmaster at globalstar.com 
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers