Crashed 10 min ago. Some context it may be totally unrelated, but the random crashing started when I added an old 1tb HDD in. It's not on the array, doesn't hold any system apps, only frigate.

 

I previously had frigate on an SSD, but I was running out of space so I decided to let the HDD hold the bulk data via mover. Here's smart for the old hdd. 

 

Anyways, at random times, unraid would crash and I'd have to force shut down and start again. Happens every few days, no pattern, different times of day. Can a bad HDD not even holding any system app data and not on the array cause unraid to crash? 

 

Around the same time I noticed that my btrfs cache pool was getting errors, but I've since switched to zfs pool and has no error. I've also ran memtest overnight and passed with 2+ passes. 

 

At this point I'm at a loss. I installed grafana to watch my ram and cpu, but right before the crash, there was no change. No cpu spiking, no ram building up. I'm still working on figuring out how to capture data on each docker container. 

ec-unraid-diagnostics-20250509-2249.zip

Edited May 10, 20251 yr by EcN

Syslog in the diags starts over after every boot, enable the syslog server and post that after a crash, but I see you are virtualizing Unraid, note that is not supported, but let's take a look at the log first.

Unraid crashed again at 5:58am. got the syslog, but there isn't much in it? I had it save in appdata. Did I miss a setting to capture more data in that log?

 

I've added grafana logs if anyone can understand it more than me: https://imgur.com/a/dHh4ava

syslog-127.0.0.1.log ec-unraid-diagnostics-20250513-0731.zip

Edited May 13, 20251 yr by EcN

That doesn't look like the Unraid syslog. Are you sure that isn't from proxmox?

 

Post a screenshot of your syslog server settings.

I see that you are virtualizing Unraid, that is not supported, recommend retesting bare metal.

 

 

Here you go: https://imgur.com/a/ooTbeje

 

I understand virtualizing is not supported and I'll keep that in mind and we can chalk it up to a virtualization issue in the end. I do want to exhaust all my options first and give as much logs/data as possible to see what troubleshooting can be done. 

 

I do plan to run bare metal eventually, but not for several months.

Edited May 13, 20251 yr by EcN

After screenshotting my syslog settings, I noticed it said "copy syslog to flash on shutdown" is set to yes. I recall this time around I was able to gracefully shutdown via proxmox. I checked my flashdrive and looks like it does have logs right before the crash. I've attached it here. 

 

Side question: Should that setting be enabled? I noticed all my diagnostics are saved here: https://imgur.com/a/qFVfVJk

syslog-previous

Edited May 13, 20251 yr by EcN

14 minutes ago, EcN said:

I noticed all my diagnostics are saved here

Is that on your flash drive?

 

15 minutes ago, EcN said:

my syslog settings, I noticed it said "copy syslog to flash on shutdown" is set to yes

Best if you don't constantly write to flash.

 

You also need to set the remote server in your syslog server settings. It can be localhost.

Yes that was my flashdrive.

Oh so I didn't properly set up the syslog server? So set the server to localhost like this? (I've also turned off copy syslog to flash on shutdown)

 

Would I go back and look in the appdata folder after the next time it crashes? Or just download diagnostics as normal? 

 

 

Edited May 13, 20251 yr by EcN

8 hours ago, EcN said:

look in the appdata folder after the next time it crashes

 

Crashed again around 14:37. Attached logs. 

syslog-10.0.1.61.log

Edited May 14, 20251 yr by EcN