I have been troubleshooting random crashes/restarts (and the associated unclean shutdowns) for months. This has led to me guessing at root cause and replacing the two NVME cache drives, all the RAM, and my UPS. More recently,  I have had the syslog on since April 9th, and had the following shutdowns:

4/10/25 @ 8:58pm

4/20/25 @ 2:03am

4/30/25 @ 3:13pm

The system is critical in the operation of our house (DNS, security cameras, HA for home automation, Plex, etc.), so I can't run it in NAS mode for 10+ days (one of the common suggestions for troubleshooting further). I'm tempted to replace the power supply, or maybe screw around with the bios, but I'm running out of ideas.

Thoughts? Diags and syslog for the past 21 days attached.

unraid-diagnostics-20250430-1735.zip syslog-192.168.1.198.log

Edited May 1, 20251 yr by bmrowe
clarifying issue

10 hours ago, bmrowe said:

I'm tempted to replace the power supply

Worth a try, especially if you have another one you can swap with, without anything logged, not really much to go on.

6 hours ago, JorgeB said:

Worth a try, especially if you have another one you can swap with, without anything logged, not really much to go on.

I'd have to buy another one. It would be a weird failure as its only 18 months old and overpowered for my build: https://www.corsair.com/us/en/p/psu/cp-9020180-na/rmx-series-rm850x-850-watt-80-plus-gold-certified-fully-modular-psu-cp-9020180-na

@JorgeB One other weird thing that I just thought of. When the system comes back up after the crash and unclean shutdown, my networking table is always jacked up. For some reason, the shim disappears on restart. Below is the routing table before and after these crashes. Any chance this is related? 

Before crash:

After crash and systems comes back online:

shim networks are created when the Docker setting "Host access to custom networks" is enabled, is it still enabled?

3 minutes ago, JorgeB said:

shim networks are created when the Docker setting "Host access to custom networks" is enabled, is it still enabled?

Yes. I can fix it by going into settings, disabled docker, and re-enabling docker.

I think that should always appear if HA is enabled, but should be unrelated to your issue.

57 minutes ago, JorgeB said:

I think that should always appear if HA is enabled, but should be unrelated to your issue.

Ok. That bug has been around awhile, was just hoping it was somehow related. (here is the history on that bug, hopefully it gets included in an update: 

 

I replaced the power supply yesterday and the issue happened today. 

 

The only hardware that hasn't been replaced at this point is:

 

CPU (i13500)

Motherboard (Gigabyte B760M)

LSI HBA card (LSI 9201-8i)

Dual Coral M.2 card

 

Temperatures on the system are never over 140F, so I don't think its related to overheating.

 

Are there instances of docker containers causing entire system crashes? I'd think that the isolated nature of a container would prevent that, but this feels more and more like its software related given the amount of hardware I have replaced..

 

 

 

9 hours ago, bmrowe said:

I'd think that the isolated nature of a container would prevent that

It's been known to happen multiple times before, one thing you can try is to boot the server in safe mode with all docker containers/VMs disabled, let it run as a basic NAS for a few days, if it still crashes it's likely a hardware problem, if it doesn't start turning on the other services one by one, including the docker containers. 

@JorgeB I think I’ve narrowed it down. 
 

I run Frigate and use a dual coral tpu m.2 plugged into a pcie adapter. This all works great and shows up in unraid. 
 

However, if I disable the TPUs from being used in Frigate (and use my cpu instead), unraid appears stable. 
 

The only thing in the logs are those ‘write back’ errors/warnings. Any idea what might be happening? 

I would recommend asking in the Frigate support thread for any known issues with that config/hardware.