| 일 | 월 | 화 | 수 | 목 | 금 | 토 |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | |
| 7 | 8 | 9 | 10 | 11 | 12 | 13 |
| 14 | 15 | 16 | 17 | 18 | 19 | 20 |
| 21 | 22 | 23 | 24 | 25 | 26 | 27 |
| 28 | 29 | 30 |
- MSCS on VMWare
- dsquery
- 터키여행
- 인문고전
- Nested VM
- windows debugging tool
- windows update
- Windows Server 2008
- iSCSI target
- Hyper-V
- failover cluster
- 작업관리자
- cluster node as Domain controller
- windows media service
- Local TempDB
- SQL Server 2008
- windbg
- ftp7.5 장애조치 클러스터
- FTP7.5
- nonpaged pool
- 안철수
- paged pool
- 클러스터
- Xperf
- Session space
- Windows Server 2016
- 프로세스 CPU 사용량
- LiveKD
- SQL Server 2012R2 FCI
- Windows Server 2016 Hyper-v Cluster
- Today
- Total
류짱:Beyond MySelf
Windows server 2003 / paged pool / Max /164M 본문
특정 서버의 paged pool의 Max가 160M 정도 인데 이 값이 기본 적으로 터미널 서버 가 설치 되어 있어도 256M정도로 할당 되어야 하는데 약 164M만 할당 되는 경우가 있습니다. (참고: Windows server 커널 메모리 값)
바로 hot-plug(=hot-add) memory 가 설치되어 있기 때문입니다.
Windows Server 2003 Kernel Version 3790 (Service Pack 2) MP (4 procs) Free x86 compatible
Product: Server, suite: Enterprise TerminalServer
Built by: 3790.srv03_sp2_gdr.090319-1204
Kernel base = 0x80800000 PsLoadedModuleList = 0x808a6ea8
Physical Memory: 4192565 ( 16770260 Kb)
PagedPool Usage: 33080 ( 132320 Kb)
PagedPool Maximum: 41984 ( 167936 Kb)
[CASE STUDY]
Why do we have around 164 MB paged pool MAX instead of 249 MB?
Windows Server 2003 Kernel Version 3790 (Service Pack 2) MP (4 procs) Free x86 compatible
Product: Server, suite: Enterprise TerminalServer
Built by: 3790.srv03_sp2_gdr.090319-1204
Kernel base = 0x80800000 PsLoadedModuleList = 0x808a6ea8
Physical Memory: 4192565 ( 16770260 Kb) <--- around 16 GB(15.99 GB)
Free System PTEs: 103914 ( 415656 Kb)
NonPagedPool Usage: 32809 ( 131236 Kb)
NonPagedPool Max: 65532 ( 262128 Kb) <--- around 256 MB
PagedPool Usage: 33080 ( 132320 Kb)
PagedPool Maximum: 41984 ( 167936 Kb) <--- around 164 MB
Kernel Full Dump, Intel x86, 32-bit Addresses, PAE=ON, 3GB=off.
When I have debugged this type of thing in the past, one cause has been that one of the machines has hot-add memory enabled in the SRAT (this is different from the DynamicMemory regkey but results in the same issue) - you can check this by dumping nt!MxPfnAllocation on machine.
3: kd> x hal!*hotplug*
80a72e50 hal!HalpMaxHotPlugMemoryAddress = union _LARGE_INTEGER 0x10`00000000
80a617d6 hal!HalpGetHotPlugMemoryInfo (struct _LOADER_PARAMETER_BLOCK *)
3: kd> ? 0x10`00000000
Evaluate expression: 68719476736 = 00000010`00000000
3: kd> x nt!MxPfnAllocation
808a66e8 nt!MxPfnAllocation = 0x1c004
It seems that hot plug memory (=hot add memory) is enabled based on the HAL global variable.
It's almost equal to nt!MxPfnAllocation.
0: kd> x nt!MxPfnAllocation
808a66e8 nt!MxPfnAllocation = 0x1c004
감사합니다.^^
