6.12.6 Unraid Crashing - help decode logs

[ginch]

I've been using Unraid for years now and currently on 6.12.6

Mobo: Supermicro X9DRH-7TF/7F/iTF/iF

CPU: Intel® Xeon® CPU E5-2695 v2 @ 2.40GHz (two of them)

128 GB RAM

 

History: Last week I noticed ads that should have been stopped by Pi-hole running in a Ubuntu VM on Unraid were coming through. Mysteriously the VM was missing from Unraid.. the whole vdisk was just ... gone. No explanation. I recreated PiHole as a docker container.

 

Updated some dockers to later versions, etc. This morning I was streaming Plex remotely and it just stopped. The player couldn't connect and see my libraries. Waited till I got home and found Unraid webUI completely unresponsive. I couldn't connect, nothing displaying on the local server monitor either. I was forced to hard shut down and bring the server back up.

 

Parity check started. I got about 3% through it after about 6 hours and streamed Plex again within my home. For whatever reason Plex stopped responding again. I noticed CPU use was high on a number of cores. I tried to run docker stats from a terminal but nothing was showing up. I stopped the parity check, then attempted to stop the array.  It's saying "Array Stopping... Stopping Services" in the bottom left status bar for at least 20 minutes and looks like it's locked up again. I still managed to pull diagnostics, please, someone point me in a direction to look?

 

Thanks!

tower-diagnostics-20240604-2215.zip

[JorgeB]

Jun  4 15:24:17 Tower kernel: macvlan_broadcast+0x10a/0x150 [macvlan]
Jun  4 15:24:17 Tower kernel: ? _raw_spin_unlock+0x14/0x29
Jun  4 15:24:17 Tower kernel: macvlan_process_broadcast+0xbc/0x12f [macvlan]

 

Macvlan call traces will usually end up crashing the server, switching to ipvlan should fix it (Settings -> Docker Settings -> Docker custom network type -> ipvlan (advanced view must be enabled, top right)), then reboot.

 

There are also issues with the cache filesystem:

 

Jun  4 21:48:58 Tower kernel: BTRFS critical (device sdt1): corrupt leaf: root=2 block=1008947806208 slot=112, unexpected item end, have 5563728467 expect 12587
Jun  4 21:48:58 Tower kernel: BTRFS info (device sdt1): leaf 1008947806208 gen 15083905 total ptrs 193 free space 5089 owner 2

 

With this type of error, recommend backing up the pool and re-formatting.

 

[ginch]

Thank you JorgeB, I appreciate your time.

 

I made the change to ipvlan and read that there were certain situations where "Host access to custom networks" or "Preserve user defined networks" should also be set to Yes. Mine are set to Disabled and No respectively. Can you advise what issues I would notice if one or both of these needed to be set to the opposite settings?

 

For the second issue of backing up and reformatting the pool, is this accomplished using the Appdata Backup plugin (which is running), or by changing the user shares from

"Cache <- Array" to just "Array", format, then set the user shares back to "Cache <- Array"? This one is strange because under the Main tab > Pool Devices there are no errors noted for either cache drive.

[JorgeB]

For the 1st question don't really know, but see if everything is working as expected, I think you don't need to change anything else.

 

For the 2nd one, if the Appdata backup plugin is copying everything important in the pool it would be enough, but probably easier to move the data to the array, and then move back, you can use this:

 

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=511923