Phantom memory problems?

[Eddington]

Hello,

I've installed OmniOS r151016 with napp-it onto a Supermicro X10SRL-F with 32GB of Kingston KVR21R15S4/8 memory sticks.

In napp-it under System -> Faults -> Errors, i'm seeing a very large number of the following errors:

Jan 14 17:54:57.0900 ereport.cpu.intel.quickpath.mem_ce <--- This is the most frequent.
Jan 14 17:55:47.0900 ereport.cpu.intel.quickpath.mem_parity <--- Pretty Infrequent
Jan 14 17:59:25.3483 ereport.cpu.intel.quickpath.mem_redundant <--- Most infrequent.

When I say frequent, there are 88,942 of these as reported by "fmdump -e | wc -l" in a 9-10 day period.

Using UBCD, I ran memtest+ and memtest for about an 8 hour period combined. Both returned zero errors.

I found a Supermicro FAQ entry that doesn't really help at Super Micro Computer, Inc. - FAQ Entry. I also see some other entries similar to mine via Google however they are not exactly the same configuration although the error is the same. And they are all rather old. I was going to open a ticket with supermicro until I ran memtest/memtest+ and both returned zero errors.

I'm not sure if I'm really having a problem or if this is a problem with OmniOS that should be reported to Illumos. Or, is there something in the BIOS (default settings) that might need to be changed for OmniOS?

Any advice / feedback would be appreciated.

TIA,

Eddington

 

[gea]

I have not seen this problem.
I would remove half of the ram and retry, then check with the other half.

As the SuperMicro info is referring to ipmi, you can disable (mostly a jumper on the board) and retry

 

[Diavuno]

is that single or dual rank?

 

[Eddington]

Thanks for you replies.

gea: I'll try the half ram / swap / other half test tomorrow and post the results.

Diavuno: Dual rank if I read the motherboard manual correctly.

 

[Eddington]

Turns out it was my fault. The memory is indeed dual channel, however when I installed it, while I read the manual, I did something entirely different. The memory modules were not installed in the correct order . I've corrected this and now the system has been up for 30 minutes without errors.

Interesting that the BIOS and Linux for a year didn't catch this or have any indication of a problem. memtest didn't indicate a problem. OmniOS however, while it worked, flagged the problem. Another example of why I've moved from Linux to OmniOS for my NAS needs .

Thanks for the help.