memory issue uncorrectable error in DIMM 1A 1B 1C 1D

[johnnymac_1]

Help. When I boot into console (Hyperterm) the systme peforms a memory self test quicly and the response is -
2 0 0x000B90 0X FFFFFFFFF011AFF74 uncorrectable error in DIMM 1A
2 0 0x000B90 0X FFFFFFFFF011AFF74 uncorrectable error in DIMM 1B
2 0 0x000B90 0X FFFFFFFFF011AFF74 uncorrectable error in DIMM 1C
2 0 0x000B90 0X FFFFFFFFF011AFF74 uncorrectable error in DIMM 1D

insufficient memory for operation
multiple memory errors detected
memory interleave operation failed
cell halt

Very new to HP-UX 11.23 Itanium. The screen does not allow me to intercept and access a Console Menu. Enter on the keybooard does interrupt the boot process. Is this a board issue since so many DIMMs are affected at once?

[Torsten.]

Without knowing the system model and other details, I can only tell you that 1 bad dimm may affect the whole quad (4 dimm group) - so likely 1 of the dimms is bad. Check the MP logs for related errors.

[johnnymac_1]

How do I check the MP Logs?

[Torsten.]

What server model???

[johnnymac_1]

rx 5670 1 Ghz

[Torsten.]

Connect to the console, type CTRL b for GSP access, select sl for system logs, then select e for errors. I assume this is the correct sequence, but your system is a bit exotic ...


Do you have some more dimms installed?

[johnnymac_1]

Those commands seemed to work. Now that I have some ability to navigate and view logs, is there a way to perform a check on the DIMMS (Active or Not Acitve)?

[Torsten.]

Without a running EFI shell you have the GSP logs only.

[johnnymac_1]

After swamping out the DIMMS for known good DIMMS in used by an identical HP-UX Model the response above is the same. Are we talking now about the system-board as oppose to memoery the problem. The error comes up pretty quick during the self test. I can get into the MP Console and see the events (which stopped recording on the 17 of May, when the problem began). The System Event Log is 86 percent full. If I look at the Live logs (reporting as they occur) they all say memory problem. Should I "clear" the log. I would like to keep the log if possible but if that would help then that's what I will do.

[Robert_Jewell]

Here is what you can do...

- Power down the server.
- Clear the event logs.
- Power on the server.
- Review the logs. In fact, you can dump all of them to the screen using the 'D' command and save it off as a file. Post that file here if you want some help reviewing them.

But, from what you say, you may be experiencing a problem with the memory controller and not the actual DIMMs.

-Bob