Does SQL clustering cause problems with publications?

Last Post 18 May 2004 12:35 AM by satya. 5 Replies.
AddThis - Bookmarking and Sharing Button Printer Friendly
  •  
  •  
  •  
  •  
  •  
Sort:
PrevPrev NextNext
You are not authorized to post a reply.
Author Messages
Bennykoh
New Member
New Member

--
14 May 2004 01:56 AM
Hi all,
I have a 4 node cluster running windows 2003 Enterprise, with 3 SQL 2000 instances.
Once of these instances, SQLMC\MC, acts as a publisher and distributor.
The subscribers include both normal SQL servers and the other 2 SQL instances in the same cluster.

A few times, I have noticed that after installing some OS security patches and failing SQLMC\MC back and forth to the default physical node, I keep getting such errors
Error: 17883, Severity: 1, State: 0 Process 60:0 (e54) UMS Context 0x11E0AF70 appears to be non-yielding on Scheduler 2
in the event viewer. My publications all become screwed up and all the subscriptions time out. I have to reboot the node once more to resolve the above error. Even then, some of my subscriptions still fail and have to be recreated once more.

I have another non-clustered SQL server that is also a publisher and distributor. However, when it is rebooted, its subscriptions do not have such problems. Does anyone know why this happens? Any help is appreciated.
Bennykoh
New Member
New Member

--
16 May 2004 05:10 PM
Hi RM,
yes, the MS Knowledge base was the first place I went to find out what error meant.
and yes I have seen the article 816840. I have already patched all my SQL servers and instances with this SQ MS03-031.
Even though the error is still more descriptive, I still do not know how this error has occured. The related article 319892 does not really help to pintpoint the possible causes either. Has anyone experienced this error and found a possible root cause?
satya
New Member
New Member

--
17 May 2004 02:23 AM
Is the h/w is High-end disk subsystems as mentioned in this KBA.
Bennykoh
New Member
New Member

--
17 May 2004 05:32 PM

Hi Satya,
Yes, my 4 node-cluster is connected to an external SAN storage. should be the equivalent of the high-end disk subsystem mentioned in the article.
I think I recall seeing this error message in addition to the original one I had mentioned previously.
However, I see these messages in the the physical node after rebooting it upon patch installation (SQLMC\MC is failed over to another node and back to this one after rebooting). I did not use any of the offending SQL commands mentioned in the article during this time.
In addition, as I have mentioned previously in this thread, I have already patched the SQL MS03-031 patch to these SQL servers prior to seeing the error messages. This is why I am stumped.
satya
New Member
New Member

--
18 May 2004 12:35 AM
Then I would sugges to report it to MS PSS for any fix.
Bennykoh
New Member
New Member

--
18 May 2004 01:55 AM
Hi Satya,
thanks for your help!
If I see this problem during the next round of OS security upgrade I will report this to Microsoft.
You are not authorized to post a reply.

Acceptable Use Policy