I have had this problem for months, but the server has been going fine. I have no clue why the log file keeps filling up or how to clear it besides restarting the server, but it fills up quickly even after I do that.

unraid-diagnostics-20241126-2121.zip

Nov 10 08:00:01 CRBUnraid mcelog: Running trigger `socket-memory-error-trigger' (reporter: sockdb_fallback)
Nov 10 08:00:01 CRBUnraid mcelog: Fallback Socket memory error count 7333 exceeded threshold: 3475821422 in 24h

 

Looks like a RAM problem, check the System Even Log in the board, there may be more info there.

Thank you! I looked, and it was a sea of red. But is there a way to tell which one is the bad one without taking them all out and constantly restarting the machine with only one stick in there at a time? Can I see which is the bad one?

On 11/27/2024 at 9:13 AM, JorgeB said:

check the System Even Log in the board

There may be more info there, including which DIMM is the problem.

I looked and keeps repeating Socket 1: CHANNEL: ? DIMM: ?

In that case I would remove one DIMM at a time and retest.

I tried that but this is an old HP z820 and I got 16 slots. I would remove one and another slot would show failing. I replace it and remove that RDIMM and a different slot from the previous 2 would show failing. I finally removed everything and started putting them in 1 by 1 and a RDIMM that was fine before says it is failing. Something different goes bad with every restart unless I make no changes then the same thing goes bad. 

Eventually I bought new higher capacity ECC ram and used only half the slots. A poor expensive Christmas present but the problems seem to go away for now.