September 25, 2003 at 10:14 am
I found the following message in my current SQL Server log. Can anyone tell me what they mean? I'm using 2000 8.00.194 on NT 5.0 SP1. There were 16 of each message and like message were grouped together.
DATESOURCEMESSAGE
410735D0Module(UMS+000035D0) (ProcessWorkRequests(class UmsWorkQueue *)+00000264)
4107382CModule(UMS+0000382C) (ThreadStartRoutine(void *)+000000BC)
77E837CDModule(KERNEL32+000037CD) (TlsSetValue+000000F0)
7800BEA1Module(MSVCRT+0000BEA1) (_beginthread+000000CE)
The databases are poorly designed and are accessed by several applications, some ColdFusion and others Java.
September 29, 2003 at 8:24 am
I'm guessing that you've got "Use NT fibers" turned on and are encountering a bug with that facility. ("UMS" stands for User Mode Scheduling, AKA fiber mode processing, AKA lightweight pooling). In EM, check the server's processor properties or just run this in QA:
sp_configure 'lightweight pooling'
Turn this off if it's on. Unless you have more than three processors that are all near 100% utilization and context switches are greater than 20000/sec, UMS will only hurt performance and destabilize the system. My opinion is that this option is for TPC-C benchmarking only. 😉
--Jonathan
--Jonathan
September 30, 2003 at 8:23 am
I checked the lightweight pooling and it's not on.
September 30, 2003 at 8:41 am
quote:
I checked the lightweight pooling and it's not on.
Then I guessed wrong.
Has the issue recurred? Did you reboot the server?
--Jonathan
--Jonathan
October 1, 2003 at 7:01 am
It's one of those things that happens and then you don't see it for a week or more. I haven't seen it since I posted the question. I'm keeping an eye on it.
Viewing 5 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply