October 9, 2008 at 8:57 am
Hello, there's something on a server I am working on where the following pattern occurs:
"Normal" behavior for about 1 minute
then
High number of disk writes for about 15 seconds where all the performance counters go up:
Physical Disk:Current Disk Queue Length
Physical Disk:Avg Disk Queue Length
SQLServer:Buffer Manager:Page Writes/sec
Looking through sysprocesses during these 15 seconds, I cannot determine what is causing it. I search by kpid > 0 and I see a bunch of sleeping processes with lastwaittype of WRITELOG (which I'm guessing is a side effect, not the cause) and a few PAGEIOLATCH_SH
What i'm trying to figure out is this:
1) Normal SQL Server behavior in which the server is so old/outdated that it can no longer handle it, or
2) A user process/setting that can be fixed to allow more continuous page writes over time (instead of cramming them in every minute).
I've checked jobs on the local box as well as linked servers. If anyone has pointers on where to look next, that would help. Asking every for every detail about the server and the in
October 9, 2008 at 4:23 pm
1. Run Sql profiler and look out for high reads.
2. Execute Sp_who2 active and look for highest IO value with lastBatchtime timestamp old enough to suggest that it was running for 10-15 seconds.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply