March 24, 2010 at 10:20 am
A customer is running SQL 2005 on VMWare ESX 4.x platform. I have another customer on a similar setup. I always observe the instance reports high signal time with the following codes
ASYNC_NETWORK_IO
CXPACKET
PAGELATCH_EX
PAGELATCH_SH
WRITELOG
On a dedicated machine with hyperthreading enabled, I can expect CXPACKET signal to get higher.
any guidance is much appreciated
March 25, 2010 at 4:40 pm
Monitored the VM host and have some additional details on the high signal times;
CXPACKET - due to shared processors between instance. My guess is OS is context switching between CPU thread schedules
PAGELATCH_XX - due to shared memory between instances. Host is configured to allow burstable instances, which lead to some sort of memory context switch delays
ASYNC_NETWORK_IO - Host is configured for truncated/ bridged network cards. Seems to affect NLB & Clustering services as well
I'm still investigating WRITELOG; I see latch delays in the perf counters but I can't think of a reason why this would translate to high signal. still researching ...
hopefully, someone else will find value with this
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply