Our core wait state is MSQL_XP.
The documented cause is:
"Caused when a task is waiting for an extended stored procedure to end. SQL Server uses this wait state to detect potential MARS application deadlocks. The wait ends when the extended stored procedure call completes."
How do I find this extended stored procedure?
Setup a trace for what?
Here are our top ten wait states.
I have no bench mark to know what this means really.
Our system shows 1200 to 1400 ms wait time most often.
Description, Wait, Total Wait
SQLTRACE_INCREMENTAL_FLUSH_SLEEP, 1068.926441, 6024.2158
MSQL_XP, 0, 250.6122539
PREEMPTIVE_OS_GETPROCADDRESS, 0, 250.6047996
ASYNC_NETWORK_IO, 0.333300003, 12.60383067
PAGEIOLATCH_EX , 1.099890011, 8.90908973
PAGEIOLATCH_SH, 0.499950005, 4.126295243
WRITELOG, 0.166650002, 2.165293104
PREEMPTIVE_OS_AUTHENTICATIONOPS, 0.166650002, 1.326802214
BACKUPTHREAD, 0, 0.927363158
BACKUPIO, 0, 0.725151497
Any clarification on the meaning and impact of these wait states would be greatly appreciated.
I've read so much about this I feel like I know less than when I started.
I know what causes them but don't know what to do or even if this level of wait state is a problem.
We use Idera SQL diagnostic manager to monitor this.
What we see is slow disk reads and writes for the data drive (D) (9ms per write) and memory paging.
We moved temp DB off of D to a solid state drive. But it didn't help at all.
I have a screen shot of the Idera Monitor I can send to anyone interested in helping me solve this performance issue.