SUMMARY : Vertias VM 3.2 reporting a FAILING disk

Hi Folks,

Thanks to the following, in order of mail receipts: 

Adrian Phipps
Shailendra Dawane
Eugene Schmidt

One said a reboot did the trick, which wasn't an option for me as this is a production server, and the other two said to use the vxedit command, to wit :

# vxedit -g rootdg set failing=false rootmir

This, I have just ran and as stated, the FAILING flags are now gone.

Thanks a bunch, original mail below.

Regards,

David S.

P.S. As regards SUN, I have already escalated the lack of response along with waiting time for anwsers (We supposedly have Gold Service......).

========= original posting =========

From: David STAPLETON [mailto:davy_stapleton@bikerider.com]


Sent: Thursday, February 12, 2004 11:56 AM
To: sunmanagers@sunmanagers.org

Subject: Vertias VM 3.2 reporting a FAILING disk


Hi folks,

I want to run this one by you to see if any of you have any
suggestions. - Apologies in advance for the long winded mail.

We recently had a lot of spurious error messages in /var/adm/messages
on one of our servers after presenting new disks from an external
storage array using JNI fibre cards (lots of forceload failures).  I
didn't think it was anything serious but logged a call with SUN as
there were so many error messages that it was difficult to see if any
"real" errors were occurring.  SUN responded and sure
enough, the
errors were harmless, the solution was to edit /etc/system and add the
line "forceload: drv/clone". 


Did this and then did some general system checking and the vxprint
command showed that the rootmir was "FAILING".  Important
to note that
there was nothing in the messages file to give any indication of an
issue, just this "FAILING" flag on the vxprint output.

As the engineer had our explorer output and the ticket was still open
I asked him about this.  He eventually replied saying that it appeared
to be just a flag setting on Veritas and had nothing to do with the
disk per se.  He added that the vxdisk output had no such error so he
was convinced it was just a flag.

At this stage I had ran an analyze of the disk, read/test and refresh
and this had shown no problems so I was inclined to believe him.  He
then told me to use the vxconfigd command with the -k option to kill
and restart the process and thus clear the flag.  Well I did this and
it had no effect, the vxprint output still had rootmir FAILING.

At this point the engineer seems to have disappeared, I jest you not.
I sent him a mail saying the command had not cleared the flag and
received nothing, but I was not too worried as he has said it was
essentially nothing.   


The core of the issue I have is that, after leaving this for a few
days, I checked the system yesterday morning and now not only does the
vxprint output still show a failing rootmir but also the vxdisk output
aswell.

Having sent another mail to the engineer and receiving no response, I
would like to know what you folks think? 


There is still nothing in the messages file to indicate an issue so
what next?

Below is the output of vxprint and vxdisk...

Regards,

David S.

========= vxprint =========
# vxprint -g rootdg
TY NAME         ASSOC        KSTATE   LENGTH   PLOFFS   STATE
TUTIL0  PUTIL0
dg rootdg       rootdg       -        -        -        -        -
-

dm rootdisk     c1t0d0s2     -        71124291 -        -        -
-
dm rootmir      c1t1d0s2     -        71124291 -        FAILING  - 


               <skipped output>


========= vxdisk =========
# vxdisk -o alldgs list
DEVICE       TYPE      DISK         GROUP        STATUS
c1t0d0s2     sliced    rootdisk     rootdg       online
c1t1d0s2     sliced    rootmir      rootdg       online failing


               <skipped output>

==========================
-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers