Correctable memory errors detected...but which DIMM?

[Greg Struthers]

I have an older DL580 running W2K Server. I have been receiving HP SIM alerts that say "A memory module has exceeded the preset threshold for correctable memory errors and should be replaced. Identify and replace the proper memory module. Newer agents exist that will provide greater support for this event and they should be installed if possible." However, the alert doesn't indicate which DIMM is causing trouble. When I look at SMH (HP System Management Homepage v2.1.2.127) and the Integrated Management Log, there is no trouble found or recorded. Server's Event Log also does not indicate any memory errors. There are 12 DIMMs installed on this machine, so not really feasible to remove them one at a time for testing (especially on a production machine).

How can I find out which DIMM is triggering this SIM alert?

[predrag81]

Hi Greg,

but which generation of DL580 you have?
And if i'm not wrong you know that problem is in A memory modul. You should check led on memory board. Or, you can:
First, update your hP software tools(ADU, IML, SMH), then clear iml log, then try to test memory from insight diagnostic and check status in SMH.

But the best way is to phisicaly access to the sercver abnd to check led status on memory modul on memory board or mainboard(depends of which generation of dl580 you have)

Maybe this could help.

[Greg Struthers]

This server is a DL580 G1. Unfortunately, it is physically located at a different site, so I can't look at it myself. I could do a full HP utilities upgrade on the server, although I was hoping for an easier and more guaranteed solution. I did try to run the Insight Diagnostics via SMH, but the only thing it showed as available to diagnose was hard drives.

[predrag81]

Yes fiend, because, you cannot test all hw component online. You must reboot server and boot with HP SS and then execute diag.

predrag

[Brian_Murdoch]

Greg,

Can you check which Compaq or HP Agent version you are running. Go into Control Panel and click on Compaq (or HP) Management Agents. The banner along the top will tell you which version you are running. There were false memory issues with older agent versions.

You may also have corresponding events in the system event log, include these event numbers in any reply please.

Regards,

Brian

[gregersenj]

There was a problem with the agents, years ago.
I'm not 100% sure, but i think it was ver. 6.30.
Problem was: If you have anything, failed or degraded in the IML: Could be a NIC with no cable.
That would caurse an entry in the windows event log, indicating memory failure, upon boot.

Solution:
Upgrade the agents.