October 7, 2003 at 7:21 am
Hi!
Few days ago a very long block chains began to happen very often.
I noticed that this happens at the moments when one or two processes
create too many locks in tempdb (up to 60000 and even more). (99% are
locks on tables syscolumns, sysindexes & sysobjects in tempdb) After
that everything nearly stops. /* to determine this I run this query:
select req_spid, count(*) as Numb from syslockinfo (nolock)
group by req_spid
order by Numb desc */
This is a sample output:
====================
req_spid,Numb
169, 58437
617, 15624
980, 404
1003, 18
58, 18
529, 18
281, 13
520, 13
573, 13
542, 13
198, 6
114, 6
79, 6
333, 6
91, 5
=====================
What can be the cause of this?
Thanks.
October 10, 2003 at 8:00 am
This was removed by the editor as SPAM
October 10, 2003 at 8:09 am
U can try to increase the no of locks . my guess would be too many processes starting at the same time, i mean jobs. SQL server has a known bug which MSFT is working on when too many jobs run at the same time that access sys tables(updates).
October 10, 2003 at 8:55 am
What has changed before this started occuring?
Don't know if this could be the problem, but do you use Autogrow on the TempDB or on the production databases?
Maybe set the 'increment' to a higher value or disable it all together and allocate a sufficient amount of disk.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply