We used to receive BUF latch errors a few months ago, along with a few other errors on our [production] server. To allievate some of the problems, we split off TempDB MDF and LDF's onto a separate drive, and installed the lastest service pack (4). Server memory did not appear to be an issue as buffer cache hits looked very good. We have not any any further problems until yesterday evening.
Yesterday, we started receiving several errors and one process ID that was blocking was killed, and remained in a continuous Rollback state. We are presently scheduling a reboot of the server at the end of the workday.
These are some of the repeating errors in the SQL Log:
Does anyone know (from these errors) what might be causing these problems and/or how to resolve them?
-Ovatvvon :-Q
Yesterday, we started receiving several errors and one process ID that was blocking was killed, and remained in a continuous Rollback state. We are presently scheduling a reboot of the server at the end of the workday.
These are some of the repeating errors in the SQL Log:
Code:
WARNING: Failed to reserve contiguous memory of Size= 65536.
WARNING: Failed to reserve contiguous memory of Size= 65536.
Buffer Distribution: Stolen=4294949362 Free=404 Procedures=242208
Inram=0 Dirty=29163 Kept=0
I/O=0, Latched=442, Other=82141
Buffer Counts: Commited=336424 Target=336424 Hashed=111746
InternalReservation=897 ExternalReservation=0 Min Free=256 Visible= 336424
Procedure Cache: TotalProcs=44500 TotalPages=242208 InUsePages=101828
Dynamic Memory Manager: Stolen=224274 OS Reserved=33448
OS Committed=33406
OS In Use=32780
Query Plan=242109 Optimizer=6
General=13868
Utilities=38 Connection=762
Global Memory Objects: Resource=8771 Locks=115
SQLCache=1661 Replication=2
LockBytes=2 ServerGlobal=24
Xact=3258
Query Memory Manager: Grants=0 Waiting=0 Maximum=189144 Available=189144
WARNING: EC 8c78e3f0, 0 waited 600 sec. on latch 8090a2c8, class MISC. Not a BUF latch.
Attempted to Kill a Process ID; stuck in continuous rollback state.
[COLOR=green]Repeats the following over and over for hours...[/color]
WARNING: EC 8c78e3f0, 0 waited 59100 sec. on latch 8090a2c8, class MISC. Not a BUF latch.
Waiting for type 0x4, current count 0xa, current owning EC 0x8D187580.
Does anyone know (from these errors) what might be causing these problems and/or how to resolve them?
-Ovatvvon :-Q