unRAID randomly locking up - soft reset sometimes fails

[dzportwood]

Recently upgraded to 6.11 but issue persists.  Some background

This was originally a dual Xeon server but I started getting RAM errors and generally wanted to upgrade to ditch nvidia GPU and just go intel quicksync for plex and others - figured this would solve the random lockups.  Did not.   Sometimes you can SSH in and shutdown/restart sometimes you can't, sometimes you can use the power button to initiate a soft shutdown other times you need to hold to force powerdown. 

 

The few things that standout to me in the logs are what appears to be kernel error, and I have seen the cache drive say there were errors.  Smart data does not seem to indicate anything wrong with cache drive (there is a known bad disk whoever in the array).  

 

I have attached support bundle. 

thehub-diagnostics-20220927-1625.zip

[JorgeB]

Enable the syslog server and post that after a crash. 

[dzportwood]

Ok - I currently use a remote syslog facility, can I just send that instead, or is there some advantage to using the integrated?

 

All-Messages-search-result.csv

Edited September 28, 2022 by dzportwood

[dzportwood]

@JorgeB - Posted above

 

[JorgeB]

8 minutes ago, dzportwood said:

Posted above

Didn't see it before, I guess you edited the post? Should always do it in a new post.

 

"2022-09-27T13:46:34.000Z","TheHub","TheHub kernel: macvlan_process_broadcast+0xbc/0x12f [macvlan]"

 

Macvlan call traces are usually the result of having dockers with a custom IP address and will 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)).

 

[dzportwood]

I did indeed edit it!

 

Curious - if I swap to ipvlan would that not make all containers share the same mac address and thusly cause DHCP issues (amongst a plethora of other drawbacks higher cpu utilization, routing configuration not available to most consumer based SoHo stuff)?  If so, that is really not a problem I can convert to ipvlan and use static assignments for all containers; albeit I rely heavily on DHCP and DNS in my network as general practice without worries of DHCP client ID's.

 

Will try this approach with static assignments ASAP and report back. 

 

Thanks for your help.

Edited October 1, 2022 by dzportwood

[dzportwood]

So far so good.  Longest uptime since the issue started was about 4 days so will report back when it reaches that.  

 

Thanks for your assistance. 

 

[dzportwood]

6 Days+ no issues, upgrading to 6.11.1 so reboot will happen but this seems to have solved the issue.

[dzportwood]

This has now returned as of update to 6.11.5 - I have verified IPVLAN is still selected.   New syslog file attached. 

 

 

All-Messages-search-result (1).csv

[JorgeB]

Unfortunately there's nothing relevant logged.

[dzportwood]

Thanks for looking, will wait for next time and see if any delta in log events.