June 9, 2004 at 10:35 am
We have Active-Active SQL Cluster and today in the morning, one node failed over to 2nd node and this is the only error log that I found.
Error: 605, Severity: 21, State: 1
Attempt to fetch logical page (3:15674) in database 'Boards' belongs to object '6226036', not to object 'Message'..
I searched the user group and only thing that I was to find is that possible database corruption and some problem with dynamic and that error caused the dump file and based on dump file, the query that we are using is dynamic cursor even though that is very complicated one. Also I did full DBCC CHECKDB and couldn’t find anything from that.
Does anyone have any idea?
Thank you in advance,
Kun
June 9, 2004 at 11:34 am
You may actually have two separate issues.
1-the cluster failed over
2-the error message
The error message may be a result of the query running while the failover was occuring.
Did you 'see' the failover happen? In other words, were you there when it happened? Do you know for sure that it occurred as a result of the query?
Did you rerun the query? If so, did you get the same error? (If not, then it's most likely something else caused the failover and the query couldn't find the proper data because the database was still failing over.)
-SQLBill
June 9, 2004 at 11:39 am
Hi SQLBill,
Yes, I actually was on sql cluster admin and watched it failed over. About 30min before sql failed and failed over, we saw hundreds of error messages from web server with that error message and getting timing out the pages. We thought it was spider(bot) that did it but it wasn't. We are running that proc about 40K times per hour and once it failed back, I don't see any problem any more.
Thanks,
Kun
June 9, 2004 at 11:43 am
Oh, one correction
the query that we are using is dynamic cursor
>>> the query that we are using is NOT dynamic cursor
June 9, 2004 at 12:01 pm
Were there any other queries running at the time?
Any maintenance being done? Backups?
This might just be one of those unexplained happenings.
-SQLBill
June 9, 2004 at 12:12 pm
No Backups or Maintenance was running. I was looking at active processes and nothing strange things were happenging at that time. The query (store proc) is called by aspx page and it's been running our site for at least 1/2 year without too much trouble except the fact is that the proc is heavy. And interesting factor is that even though I was getting those error log, over 95% users were able to access the page. Even I didn't see that from web page.
Very strange..
June 9, 2004 at 1:43 pm
I think this is going to end up in the 'Unsolvable Glitch' folder.
-SQLBill
June 9, 2004 at 1:46 pm
It sounds like it. Thank you for the input
June 9, 2004 at 3:01 pm
You might want to verify your backup. Below is the link to a simillar issue
http://www.sqlservercentral.com/forums/shwmessage.aspx?forumid=9&messageid=115255
June 9, 2004 at 3:06 pm
Hi sa24,
It wasn't related to database backup or database status makred as suspect. I don't think it is related to my case but thank you for the link.
August 4, 2004 at 7:43 am
I dont know if this is of any help.. but still
We had the same kind of problem on sql 2000 server after we installed OS SP3 on Win 2000. After talking to MS for months and doing lots of hardware checks we found that the in SP3 the write cache for disks was enabled. When we manually disabled the write cache the problem stopped happening.
Here is how you can disable the write cache :
Please let me know if that helps.
Thanks
September 8, 2005 at 9:05 am
Check out this link
http://support.microsoft.com/default.aspx?scid=kb;en-us;826433
It details the problem.
Thanks
Viewing 12 posts - 1 through 11 (of 11 total)
You must be logged in to reply to this topic. Login to reply