July 2, 2002 at 7:56 am
We are running SQL 97. Our Stolen Page coounter keeps rising until we get timeouts. The CPU utilization is not high. This happens even if we only have two connections. There is no other software running on the databse server. Any help would be appreciated
July 2, 2002 at 9:41 am
What is SQL 97? SQL Server 7.0?
From BOL:Number of pages used for miscellaneous server purposes (including procedure cache).
This is the def for the stolen pages. If you have only 2 connections, what are they doing? Also, are there any jobs or other processes running in SQL?
As this rises, can you capture a few "sp_who2" outputs and see if anything is running?
Steve Jones
July 2, 2002 at 9:49 am
Sorry SQL Server 7
The two connections are using our custom VB6 app. This involves stored procedure calls and data inserts. We have as many as 30 connections without experiencing this kind of performance degradation.
No other jobs or processes running in SQL
Used Enterprise manager instead of "sp_who"
We also have users connecting through Citrix. They are usually the first to notice a slow down. The afore mentioned two connections were NOT Citrix users
Thanks
July 2, 2002 at 10:13 am
Don't really see anything in Technet or BOL related to this.
How much memory on the server? Does available memory decrease? Does SQL memory increase?
Run profiler and verify if anything else is happening.
Steve Jones
July 2, 2002 at 10:38 am
From PerfMon: Number of pages used for miscellaneous server purposes (including procedure cache).
In addition to Steve's questions... Do you have a lot of stored procedures? Are they complex? Are any jobs running through SQL Server Agent?
K. Brian Kelley
http://www.sqlservercentral.com/columnists/bkelley/
K. Brian Kelley
@kbriankelley
July 2, 2002 at 10:53 am
We have 2 Gigs of memory
Replication is running through SQl Server Agent but we have a remote distibutor.
Monitoring Available and SQl memory right now.
We do have a lot of stored procedures but they are fairly simple. Our tables are large up to five million rows.
We do use triggers to poulate out two history tables
Thanks
The strange thing is that we have been functioning fine until one week ago.
We did a large data reduction and everything went back to normal until today.
July 2, 2002 at 12:14 pm
Have you run a backup since then?
Steve Jones
July 2, 2002 at 12:17 pm
What do the numbers on your buffer cache hit ratio and pages/sec look like?
K. Brian Kelley
http://www.sqlservercentral.com/columnists/bkelley/
K. Brian Kelley
@kbriankelley
Viewing 8 posts - 1 through 8 (of 8 total)
You must be logged in to reply to this topic. Login to reply