CACHE MEMORY alert from BMC

[HP UNIX Professionals]

Hi Experts,

I am getting CACH MEMORY alert daily form HPUX box through BMC, But i am not able to see any abnormalities is cstm or syslog.

I am posted the CSTM and memory information please advice anybody know the solution for this alert.

BMC Alert -Impact[MEDIUM] Device[Server123:HP:MEMORY:MEMRCache:MEMORY] NewState[CRITICAL] Info[PATROL Agent CRITICA

=========================================
Memory Information:
physical page size = 4096 bytes, logical page size = 4096 bytes
Physical: 14678016 Kbytes, lockable: 13484728 Kbytes, available: 13008636 Kbytes

===========================================

CSTM

-- Information Tool Log for MEMORY on path 8 --

Log creation time: Mon Oct 18 12:15:40 2010

Hardware path: 8


Basic Memory Description

Module Type: MEMORY
Total Configured Memory : 14336 MB
Page Size: 4096 Bytes

Memory interleaving is supported on this machine and is ON.

Memory Board Inventory

DIMM Slot Size (MB)
--------- ---------
0A 512
0B 512
0C 512
0D 512
1A 512
1B 512
1C 512
1D 512
2A 512
2B 512
2C 512
2D 512
3A 2048
3B 2048
3C 2048
3D 2048
--------- ---------
System Total (MB): 14336

Memory Error Log Summary

The memory error log is empty.

Page Deallocation Table (PDT)

==============================================

root@server123:/# swapinfo
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 12288000 0 12288000 0% 0 - 1 /dev/vg00/lvol2
dev 16384000 0 16384000 0% 0 - 1 /dev/vg00/swap02
reserve - 2649692 -2649692
memory 11351640 7581600 3770040 67%



Thanks in advance

VINAY

[Steven E. Protter]

Shalom,

I think the error message is not referring to system memory which you have checked and found to be okay.

cache may refer to CPU cache, which is a little area of very fast memory designed to speed up CPU operations.

Use cstm,mstm or xstm to test the CPU's. Contact hardware support and see if the message is important or I am wrong.

SEP

[Torsten.]

I'm not sure from where you get this message, but in case of a cache error there should be a log entry in MP system logs.

Check this.

[Jim Walls]

Perhaps you should ask your BMC support to explain, exactly, what this alert means - and how (what procedure is used) it is detected.

[Torsten.]

BMC is the baseboard management controller - its a part of server hardware.

For exactly this reason the MP should log related events.

[Dennis Handly]

>Torsten: BMC is the baseboard management controller

Are you sure this message is from that BMC?
google finds "create_new_parameter: MEMRCache" hiding away in some bmc.com page.
The two hits on hp.com refer to bmc.com too.

[Torsten.]

See my first post.

[Duncan Edmonstone · Accepted solution]

A google search turned up this:

MEMRCache: Displays the percentage of read misses in the buffer cache.

Found in this doc here:

http://www.scribd.com/doc/24178426/Bmc-Patrol-Unix

So this is nothing to do with physical memory - it's related to the number of hits in the file system buffer cache for reads. This sounds like one of those arbitrary settings that BMC Patrol has - I would look at the thresholds you have set and adjust them appropriately.

HTH

Duncan