Hi there,

last week my server crashed, it happens, so I didn't think much of it. But just two days later it crashed again, and this shouldn't happen :(

Both times it rebooted, did the parity check, and it was fine (first crash gave an XFS error, but it could be fixed easily, thanks again JorgeB). I can't recall to have done anything special during those crash times, I only noticed from the notification+parity check.

The 2nd crash happened on the 17th October, 15:00 h. I wanted to make this post earlier, but was away for the weekend so I wouldn've had a chance to react to any responses :(

I had a look at the syslog I am writing constantly, but it looks like it's not at all helpful for that timeframe. I attached the syslog from beginning of the day to short after the crash; to me, it looks like nothing is logged for the exact moment, just the reboot-sequence.

I am aware the usual culprit is RAM, so I did a memtest (the one I can boot from the unRAID Stick). I let it run 12hours, and got the big "pass" screen (see attached images). Note that this is ECC (unbuffered) RAM, but the mainboard is not capable of ECC.

Also I attached the diagnostics I downloaded shortly after the crash (the moment I noticed it actually did crash), so around 16:22h on the 17th of October.

Any ideas what it could have been? Is disk7, that had the XFS error after the 1st crash, maybe broken somehow (even through it shows fine)?

Best Regards

lunas-diagnostics-20251017-1622.zip syslog_251017_till1515.txt

There's nothing relevant logged; this usually means a hardware issue, since memtest is only definitive if it finds errors, and you have an unusual RAM configuration (though valid) I would recommend running the server with just the 32GB stick, if still crashes, try running it with the 8GB pair. That will basically rule out a RAM problem.

Hi there Jorge,

before I switched to this Intel system in January, I had an AMD system which is where the ECC memory comes from. DDR4 was/is kinda expensive, so I just kept that one. I'm never sure how much memory I actually need, by now I am not using VMs much so 32GB is probably fine. I will try to carve them out in the next few days to get an idea.
It's a shame I don't have a 2nd system where I could run memtest for 24h+ for the DIMM individually while the server still runs, that would be really handy.

Best,

Rick