March 29, 2006 at 10:03 am
Hi all. I've read about a million articles on Google adn Microsoft about possible causes but nobody seems to have the definitive answer. Twice a week I run a maintenance plan to update statistics (100%). Every one in a while, about every 6-8 weeks, I'll come in in the morning and someone will tell me they cannot connect to the database. If it's Tuesday or Thursday morning, I know what has happened:
The Scheduler 3 appears to be hung. SPID 77, ECID 0, UMS Context 0x0686EFC8.
Error: 17883, Severity: 1, State: 0
My only recourse is to the kill the job and then reboot the server. It's a dual-processor, hyperthreading, running SQL 2000, sp3. Some articles state it's fixed in sp4, some say it's not. Some state it's the hyperthreading and to reduce the amount that SQL uses to 2 (the actual physical number of CPUs). I've also read it could be the "AUTOGROW" feature (which is turned on) but I know my files don't grow that frequently. I also know there is, supposedly, a hotfix to resolve this but before I do aanything along those lines, I figured I'd pose the question to you: Any of the experts exprience this and, if so, are you willing to share what you did to resolve the problem??
Terry
March 30, 2006 at 1:04 am
Terry,
What is the build of SQL Server you are running?
Please run select @@version and check whether its 8.00.760 or 8.00.818.
If its 818, you will see the dump file generated which can be used to find the root cause.
If you are on 760, there is no way you can find the cause of this. SP4 is definity a good choice to avoid known issues of 17883.
Are you sure that this happens when you run update stats? As per Error Message SPID 77 caused this. Did you get chance to see the SPID of the job which was running for update stats?
Feel free to ping me at bmlakhani@yahoo.com
March 30, 2006 at 8:38 am
It is 8.00.760 I can also raise this error when I restore a database in the development environment. The servers are identical, dual CPU and hyperthreading. It does run to completion when doing the restore but as stated, on occasion with the update statistics, it becomes a runaway job that I ultimately have to kill and reboot. What will SP4 give me, beside the dump? Will SP4 and the hotfix for AWE eliminate this issue or just give me better error reporting?
Terry
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply