Re: 11g R2 RAC node eviction log details missing

From: Sanjay Mishra <smishra_97_at_yahoo.com>
Date: Tue, 2 Jul 2013 12:58:32 -0700 (PDT)
Message-ID: <1372795112.54954.YahooMailNeo_at_web122101.mail.ne1.yahoo.com>

Can someone help in interpreting the following lines in cssd log

[CSSD][1240823911]clssgmUnregisterShared: Same group share client 1 (0xbb65ba0), grp DG_DATA, member 4
[CSSD][1240823911]clssgmTermShare: (0xbd56a30) local grock DG_DATA member 4 type 1
[CSSD][1240823911]clssgmUnreferenceMember: local grock DG_DATA member 4 refcount is 17
[CSSD][1240823911]clssgmDiscEndpcl: gipcDestroy 0xda420cb6
[CSSD][1240823911]clssgmDeadProc: proc 0xb346a20
[CSSD][1240823911]clssgmDestroyProc: cleaning up proc(0xb346a20) con(0xda420c7e) skgpid 19194 ospid 19194 with 0 clients, refcount 0
[CSSD][1240823911]clssgmDiscEndpcl: gipcDestroy 0xda420c7e
[CSSD][1240823911]clssnmSendingThread: sending status msg to all nodes

TIA Sanjay

---

 From: Justin Mungal <justin_at_n0de.ws> To: smishra_97_at_yahoo.com
Cc: Guillermo Alan Bort <cicciuxdba_at_gmail.com>; oracle-l <oracle-l_at_freelists.org> Sent: Tuesday, July 2, 2013 11:05 AM
Subject: Re: 11g R2 RAC node eviction log details missing  

If there are zero details in any of the CRS related logs about fencing or eviction or anything of that nature, then it sounds like the servers were not evicted by the clusterware. Rather, they hit a hardware/driver fault and restarted. I've seen this happen during storage maintenance work as well. It's tough to ever find a root cause because storage people never want to admit that anything could have gone wrong. If you're using Powerpath or something similar, you may want to check to see if updates are available, or if there are bugs in the release you're using, etc... Good luck.

On Tue, Jul 2, 2013 at 9:53 AM, Sanjay Mishra <smishra_97_at_yahoo.com> wrote:

Alan
> 
>I am suprised that all 4 node rebooted and none of them has the details in the logs. 11g R2 RAC was improved to avoid rebooting and try to Restart instead of rebooting unless in severe condition it will go for reboot. Even in this case all 5 node goes to reboot almost same time.
> 
>Sanjay
>
>________________________________
> From: Guillermo Alan Bort <cicciuxdba_at_gmail.com>
>To: Sanjay Mishra <smishra_97_at_yahoo.com>
>Cc: oracle-l <oracle-l_at_freelists.org>
>Sent: Tuesday, July 2, 2013 9:41 AM
>Subject: Re: 11g R2 RAC node eviction log details missing
>
>
>
>
>Check out the databases' alert logs, if there were scsi errors chances are you lost access to datafiles or controlfiles at some point and that will be logged there.
>
>
>Other than that, what more evidence do they need? they are working on the storage, you get SCSI errors and the CRS reboots the node. This is clearly a node loosing access to the Voting disks or OCR
>
>hth
>
>
>
>Alan.-
>--
>http://www.freelists.org/webpage/oracle-l
>
>
>

--
http://www.freelists.org/webpage/oracle-l

Received on Tue Jul 02 2013 - 21:58:32 CEST