first unclean shutdowns now the Array won´t start anymore

[Reclaim]

Hey,

 

I've been struggling with unclean shutdowns for some time now.

Unraid get´s completely unresponsive, no webui, no pinging, no ssh, no Caps Lock light on the keyboard.

Only way to restore it, is to cut power.

I had to cut power yesterday, after that i enabled Syslog. After 18 hours (12.Nov 07:71) it crashes again.

 

In syslog i see:

Nov 12 07:51:31 unRAID kernel: kernel BUG at arch/x86/kvm/mmu/tdp_mmu.c:560!

 

After this unclean restart i changed the docker net from macvlan to ipvlan. Since then the Array does not start anymore.

Stuck at "Array Starting".

 

Can someone help?

unraid-diagnostics-20231112-1414.zip syslog

[Reclaim]

Update:

I was waiting 1 hour in "Array Starting". Tryed the Power Button, it says "shutting down" on console but nothing happens. Then I designated to cut power again.

Now booted in Safe Mode and Docker, VM´s disabled I was able to start the Array again.

Here is the new Syslogsyslog (1)

 

I see

Nov 12 14:28:43 unRAID kernel: WARNING: CPU: 4 PID: 40 at kernel/kthread.c:141 free_kthread_struct+0x1e/0x43

after i started the Array last time.

 

 

[Reclaim]

Update: 

Now I can´t start the array even when booted in Safe Mode.

It stuck at mounting the zfs cache.

 

This is in the Syslog:

Nov 12 23:53:03 unRAID kernel: PANIC at dnode_sync.c:301:free_children()

 

 

syslog (3)

[JorgeB]

There's likely an underlying hardware issue, but you can try importing the pool read-only:

 

zpool import -o readonly=on cache

 

If it works backup and re-format the pool

[Reclaim]

Thanks, i could import the pool as read-only. 

After the backup i will reformat the pool.

 

Do you think it's still a hardware problem? How can i determination which component has the fault or have you a guess?

[JorgeB]

There are a lot of call traces logged, also the pool getting corrupt may be a symptom, start by running memtest.

[Reclaim]

Update: I only formatted the Cache Pool, since then the Server is stable. Uptime now 4.5 Days.

Is it possible that a out of memory error caused the corruption?

 

One or two times i had "Out of memory errors detected on your server" on Notifications.

[JorgeB]

20 hours ago, Reclaim said:

Is it possible that a out of memory error caused the corruption?

Seems unlikely, but won't say it's not possible.