June 1, 2004 at 3:27 pm
Have an 8.00.818 sp3 SQL Server running on W2K sp4. The user databases are really small, <600mb, running on a Dell server 2cpu 1gb ram, on local Raid array.
Have installed security patch MS03-031.
Getting sporadic messages:
Error: 17883, Severity: 1, State: 0
Process 0:0 (c68) UMS Context 0x0D1536E8 appears to be non-yielding on Scheduler 0.
This database is not in production yet, has no activity; usually the timing of the 17883 is around the time of database backups. Backups are done through Veritas SQL agent.
Have a few other similar SQL servers with agent, not getting this error on any others.
Any thoughts?
thanks...
June 4, 2004 at 8:00 am
This was removed by the editor as SPAM
June 23, 2004 at 7:37 am
I have the same error .
I have found this KB in microsoft support that explain the BUG:
KB810885
High-end disk subsystem may experience error 17883
and related fix.
June 28, 2004 at 12:20 pm
August 3, 2004 at 3:48 pm
This Error message is the cause of using SQL Server with High End Disk Subsystem like EMC. Please monitor your disks and organize the databases to be load balanced so that one disk is not being hit.
August 3, 2004 at 4:01 pm
I don't believe that is the case here.
We're using a Dell Raid configuration, that would hardly be classed as high end like an EMC. The databases aren't doing any work yet, it's not in production yet - just being backed up at night, so there is no dynamic growth of data files (like the Microsoft article mentions).
I found another posting somewhere else (can't remember the site) where they had same problem with Dell hardware, theirs was also fixed by updating the firmware. Their hardware was similar to mine.
Haven't had any re-occurrence of the problem since the firmware was upgraded...
January 26, 2005 at 8:48 am
I'm having the exact same problem with an MSDE instance running on an HP ML330 G3 with 2 2.4Ghz CPU's (with hyperthreading) and 1.5 GB of RAM. There is one RAID 1 on the server configured at the hardware level utilizing two 72 GB drives. The drive is configured with an 8 GB OS partition (drive C) and a 65 GB data partition (Drive E). The swap file resides on Drive E. MSDE was installed on Drive E and the single user database (less than 100 MB) also resided on Drive E.
I am seeing this error at various times throughout the day just about every day on this server. It has other applications running on it, but the CPU utilization is low because we have Patrol running on this machine and it is not alerting on high CPU usage. We also have BMC's Best/1 product and the charts show low resource usage. So I cannot figure out what is causing the problem.
I installed SP3a a while back and that's when the errors began to appear. We later applies MS03-031 taking the server to version 8.00.818 and it didn't resolve the problem. I later applied a hotfix for a separate issue taking the version up to 8.00.873 and the problem continues to occur.
Anyone have any ideas? I am thinking of opening a case with the SQL Team at Microsoft, but I don't expect to get very far with them.
Thanks,
-Tony-
-Tony Clark-
January 26, 2005 at 12:51 pm
The thread (scheduler) spawned by one of your two processors for an execution context of your process (backup in this case) is unable to release control of some resource, mostly the memory pool in this case. Backup is an I/O intensive operation and once the context has been established for the process, the thread should get idle till I/O (in this case, writing to the disk) gets completed. But it is not.
If it's a good RAM, well and good. Else, RAM is one of the issues here.
March 6, 2006 at 9:26 am
I'm getting the same error, i.e.
Error: 17883, Severity: 1, State: 0
Process 0:0 (c68) UMS Context 0x0D1536E8 appears to be non-yielding on Scheduler 0.
plus
SQL Server has encountered 8 occurrence(s) of IO requests taking longer than 15 seconds to complete on file [...] (11). The OS file handle is 0x00000470. The offset of the latest long IO is: 0x000002b0a7b400
and finally last week SQL restarted itself twice.
The SQL server is SP4, and it's running on a Dell array PERC 3/Di v2.8-0
Any ideas?
March 6, 2006 at 1:04 pm
Hello Folks,
The 17883 errors are very simply Sql Server's way of telling you that some thread that was executing did not release the cpu for more than 60 seconds. If you were to print out the Sql Books Online and count the pages that is about how many reasons there are for getting this error.
In many cases it is related to disk IO and specifically firmware or other driver versions. In others the underlying issue is network IO.
The thing to do is to find out what the underlying cause of the thread not releasing the cpu is. A couple of folks on this thread resolved the issue by updating the Dell disk controller firmware. We have seen similar problems on every major hardware manufacturer's products.
One thing you can do is upgrade to Sql SP4 plus the latest rollup hotfix. There was a lot of work put in to getting Sql Server to report the underlying error instead of the more generic 17883 in SP4. There are additional fixes in the rollup as well.
Hope this helps.
If you have additional questions about this reply directly to me. Unfortunately, I don't get to read the forums every day.
Richard L. Dawson
Microsoft Sql Server DBA/Data Architect
I can like a person. People are arrogant, ignorant, idiotic, irritating and mostly just plain annoying.
May 6, 2006 at 3:55 am
Viewing 11 posts - 1 through 10 (of 10 total)
You must be logged in to reply to this topic. Login to reply