deadlocks happening regularly

MediaBin 4.5.3
Windows 2003 EE SP1

We continue to see the following 2 deadlocks happening hundreds of times today:

02/21/08 13:09:39 Error with SQL: select jobid, type, priority, state, cancelrequested, paused, userid, username,datediff(second, '1/1/1970', queuetime), taskname, assetid, filename, serverid from Jobs where jobid = ?, {A3886D05-D2EB-4260-B1DB-69C163642687}
02/21/08 13:09:39 Transaction (Process ID 62) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction.

02/21/08 13:02:53 Error with SQL: update LRUCache set path=?, filesize=?, lastAccessed=getdate() where imageId=? and imageRevision=? and taskId=? and taskRevision=? and crc=?, '\\cwapdam01\MBDataDirectory\Cache', 13402, {5905FD4C-6448-4122-B9A0-70D338C56061}, 2, {A5EE6DB0-A5A3-4680-A2AC-4DB6B4664B21}, 0, '224840112'
02/21/08 13:02:53 Transaction (Process ID 72) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction.

Deadlocks should not occur this often. I can't imagine this is normal. Anyone have any insights into this?

Find more posts tagged with

Comments

msrinivas

How big is the DB and what is your reindexing schedule?

Bill Klish

DB is about 50 GB. Reindexing every sunday night (assuming you mean indexing via the enterprise manager app). Nightly we are doing SQL Server index maintenance cleanup so we don't have fragmentation.

Migrateduser

We have been getting deadlocks too. In fact, it looks exactly like yours. Keep us in the loop with the fix.

02/22/08 06:08:04 Error with SQL: select jobid, type, priority, state, cancelrequested, paused, userid, username,datediff(second, '1/1/1970', queuetime), taskname, assetid, filename, serverid from Jobs where jobid = ?, {7398C203-05BA-4DDB-906D-036845518BD7}
02/22/08 06:08:04 Transaction (Process ID 59) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction.
State:40001,Native:1205,Origin:[Microsoft][ODBC SQL Server Driver][SQL Server]
02/22/08 06:08:04 GetJob({7398C203-05BA-4DDB-906D-036845518BD7}): Succeeded in retry of deadlocked transaction (retries:1).

lyman

Unless you are getting messages that the deadlock doe snot succeed in retries, the warnings are benign. SQL server automatically handles locking and in some cases makes assumptions that lead to problems. The job table in particular has a high level of concurrent access due to its nature. There is nothing to be concerned about unless the log indicates that the transaction never succeeds.

Cheers,

Lyman Hurd