January 26, 2009 at 5:36 am
Hi all you gurus out there,
Maybe you are able to help me with the following phenomenon.
We have a database maintenance job running daily which includes an update of the statistics of all user tables. It is always the update for the same table which mostly (not always!) ends with an error message of a problem in tempdb, not the user database (I will try to translate as best as I can from German to English, but it may be that the resulting wording differs from the original English message):
[font="Courier New"]During a read operation at offset 0x00000001646000 in the file 'C:\\Programme\\Microsoft SQL Server\\MSSQL.1\\MSSQL\\DATA\\tempdb.mdf' the operating system has returned the error Outdated Page (a Log Sequence Number (LSN) (x:y:z) has been returned to SQL Server at a page read, which is older than the last written LSN (0:0:0)).
(x y and z are integers and may vary.)[/font]
I tried DBCC CHECKDB (tempdb) and DBCC CHECKDB (MyUserDB) frequently, but neither detects any inconsistencies in the databases, which leaves me rather confused. I would expect either DBCC CHECKDB to come up with some inconsistency or the maintenance job to run flawlessly.
Does anyone have an idea, what may be wrong here?
Best regards,
Dietmar Weickert.
January 29, 2009 at 3:39 am
Hi all,
I can hardly believe it: is there really nobody, who has any answer? At least a hint?
Best regards,
Dietmar Weickert.
Best regards,
Dietmar Weickert.
January 29, 2009 at 4:19 am
Looks like a stale read or a lost write, usually caused by a mis-behaving disk cache. Check your hardware, see if there are any errors in any logs
I don't like the last written LSN though. I don't think that should ever be 0 for any database.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
March 13, 2009 at 12:14 pm
Dear Gail,
After having tried for a long period to find any HW errors without success we decided to give up for the moment and to watch the system as closely as possible. Furthermore we will create more backups than before.
Thank you for your assistance!
Best regards,
Dietmar Weickert.
Best regards,
Dietmar Weickert.
March 19, 2009 at 12:16 pm
1. Have you tried recreating the TempDB?
2. Have you looked at excessive Disk fragmentation
It is strange that it is happening on the TempDB as the TempDB is more just a hashing ground. I will be keen on understanding more about your setup.
- Hardware
- OS
- SQL Version
- Location of windows pagefile
- Default Size of TempDB
- SQL Service restart schedule if any
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply