Prev: Moving DFS Staging path from default location?
Next: WS2008 COM+ problem
From: Nick on 5 Jan 2009 17:46
I have a HP DL320s server. Recently we've been having a out of paged pool
memory issue. I can usually go about 5 days before users cannot access large
files on this server, my Backups start to fail, and/or terminal services
won't let in. I've ran poolmon and memtriage and the top resource hogs when
this happens are:

Tag Type Allocs Frees Diff Bytes Per
Alloc Mapped_Driver
NtFB Paged 26637 24847 1790 79695280
44522 [ntfs][ntfs.sys - BitmpSup.c]
MmSt Paged 940411 938072 2339 39947000
17078 [nt!mm - Mm section object prototype ptes]
UlHT Paged 1 0 1 4198400
4198400 [http.sys - Hash Table]
NtfF Paged 132529 130071 2458 2300688
936 [ntfs.sys - FCB_INDEX]
CMAl Paged 1581 1330 251 1028096
4096 [nt!cm -internal registry memory allocator pool tag]


When the system first boots up, the same 2 drivers use:
MmSt Paged 31498 22223 9275 14552600 1569
NtFB Paged 602 547 55 1978472 35972

Someone else having a very similar problem said the issue was due to the VSS
shadow copy service. I've already installed the VSS roll-up fix and it
didn't help at all. I've checked all the processes in task manager, nothing
unusual or using over 100k of paged or nonpaged memory, no excessive handles
open on any processes (System has about 3400, but that seems to be normal).
This server isn't heavily used either, maybe 50 open files at any time. I am
using DFS replication for one directory, but disabling that service made no
difference.

Has anyone else encountered this problem? I'm at a loss, because the only
drivers reporting the memory leak is Microsoft (3rd party drivers seem to be
fine, not seeing any memory rise indiscrepencies there).
 | 
Pages: 1
Prev: Moving DFS Staging path from default location?
Next: WS2008 COM+ problem