Hi all,
I've having problems on a SQL 2008 R2 (10.50.4266) Enterprise server, 32 cores and 512GB RAM. Since we installed the latest CU patches we have been getting a lot of undetected deadlocks. By this I mean I can see that SPID x is blocked by SPID y and SPID y is blocked by SPID x and has been for minutes (if not hours).
I can also easily recreate a deadlock in temdb using 2 spids and it is not detected - It should always be detected within 5 seconds.
Quest spotlight also has a bunch of SPID's running all with a waittype of DEADLOCK_ENUM_MUTEX. This waittype (according to BOL/MSDN) is because only a single deadlock monitor thread can run at a time. Oddly none of these SPID's show as blocked, so I cannot tell if it is a Quest SPID blocking or the internal SQL deadlock monitor blocking. I suspect the latter. Something must be hogging the resource that cause the DEADLOCK_ENUM_MUTEX wait, but I can't see what.
I know disabling Quest would tell us if it was the cause but that means I lose all monitoring over Christmas when we are on skelton staff.
I was wondering if anyone has seen anything similar - Googling the waittype mentioned returns very little information.
Thanks
Simon