topic Re: Memory problem in Operating System - HP-UX

Memory problem

Morongwa — Mon, 01 Mar 2004 04:16:45 GMT

Goodday,

Could someone advice on memory problem, the error below was sent to me via cron, but swapinfo -m shows that memory is 49% used. Please help.
/system/events/memory/8 is >= 3.
Its current value is CRITICAL(5).
Event data from monitor:
Event Time..........: Mon Mar 1 02:59:08 2004
Severity............: CRITICAL
Monitor.............: dm_memory
Event #.............: 4500

Re: Memory problem

Robert Binkhorst — Mon, 01 Mar 2004 04:20:08 GMT

Hi,

This means that something is seriously wrong with one of your memory modules. That email should also contain a link or a command to execute. Do that, it'll tell you what's wrong exactly. You'll probably need to log a hardware call.

HTH,

Robert

Re: Memory problem

Hoefnix — Mon, 01 Mar 2004 04:24:51 GMT

It means that you have a correctable Single BIT error in a memory bank.

Use STM to figure out wich bank and call HP to replace it.
See:
http://docs.hp.com/hpux/onlinedocs/diag/ems/dm_memory.htm

HTH,
Peter

Re: Memory problem

Morongwa — Mon, 01 Mar 2004 04:26:18 GMT

Thanks Robert for your quick reply,

The msg that was sent had no suggetion for a command to execute. I am new to HP-UX and will appreciate it if you paste me the relevent commands to trableshoot this problem.

Thanks in advance.
Morongwa

Re: Memory problem

T G Manikandan — Mon, 01 Mar 2004 04:35:24 GMT

What OS version and STM version you are running.

Re: Memory problem

Hoefnix — Mon, 01 Mar 2004 04:35:28 GMT

Run the stm from the commandline.

(or xstm for the GUI)
Select the memory item (last entry on the first page (after the OK prompt you get after start stm at command line))
Then start from menu(using tab) the infromation->run feature and see which back has Single Bit Errors.

HTH,
Peter

Re: Memory problem

Robert Binkhorst — Mon, 01 Mar 2004 04:37:13 GMT

Hi,

As this is hardware issue, you'll need to log a call with your service provider. They'll need to replace it for you.

Peter already told you the command to use, for a full list of features read the man page, a short list of commands follows:
- cstm
- map (look for the memory banks)
- select device
- infolog

cstm is the command line Support Tool Manager from HP.

Good luck,

Robert

Re: Memory problem

T G Manikandan — Mon, 01 Mar 2004 04:41:20 GMT

1.Make sure you have the latest STM patches.

2.Shutdown the machine,re-seat the memory DIMM modules.

Re: Memory problem

T G Manikandan — Mon, 01 Mar 2004 04:43:56 GMT

check the troubleshooting doc

Re: Memory problem

john korterman — Mon, 01 Mar 2004 04:48:18 GMT

Hi,
the message probably came from a cron job that failed because of this hardware memory error. The error should be recorded in the syslog.
If the other suggestions do not give the expected result, then execute this:
# more /var/adm/syslog/syslog.log
and press the space bar until you reach the specified time, Mon Mar 1 02:59:08 2004.
The log should hold details about a command to execute in order to obtain more information. If that is not understandable, then please post the details.

regards,
John K.