[Solved]Unraid 6.9.2 Total crash unresponsive, WebGUI half works

[skippy4hammypc]

I've Had UNRAID for some time now, but as of late(past week or two) almost like clockwork my server takes a dump each day. in the past i have made no effort to dig down to the root cause, however now it is becoming a nuisance. I will attach my syslog's that I hope have an answer. the server will be running and doing things according to "top" from the webterminal and direct terminal, however I can not get the webgui to reboot/shutdown after 30min I get restless and snuff out the poor thing off, wait a few moments and power it back on. I've tried to stop docker and the array from the webgui to no luck, even tried "powerdown" from direct console from the server and again 30min passes and I kill the power myself. it would seem as though when the server crashes in this manor it takes the entire system. If anyone can glean information to help my situation, I would be eternally grateful.

syslog8.1.2021 jarvlos-diagnostics-20210810-1953.zip

[ChatNoir]

Are you using custom IPs for docker containers ?

 

I am not so good with this but I see mention of MACVLAN in the call trace and 6.9 is known to have issues with those.

[JorgeB]

6 hours ago, ChatNoir said:

but I see mention of MACVLAN in the call trace

Yes, looks like that common issue, upgrading to v6.10 might help, or see here:

 

https://forums.unraid.net/topic/70529-650-call-traces-when-assigning-ip-address-to-docker-containers/

See also here:

https://forums.unraid.net/bug-reports/stable-releases/690691-kernel-panic-due-to-netfilter-nf_nat_setup_info-docker-static-ip-macvlan-r1356/

 

[skippy4hammypc]

So as a precaution (likely not needed) I ran the included memtest in unraid. To no ones surprise it completed 1 pass, I know that one set of data does not make a trend but can be a better start than not starting anywhere. However I did have a docker with a static IP, though the static mapping is not "required" in the way that I "Have" to have that IP, moreover I do need that docker to run on the interface that corresponds with that subnet due to large network transfers on the other interface. It is one of my goals to get myself a smart switch and use VLans to further segment the network and have more control with systems that I think should run in the background almost invisible, save for system updates (we all love 'em)🙄. But as of yet I don't have one. The information i gathered from the response's of ChatNoir and JorgeB, and links, I have made some changes to my picky docker container to use the interface I want it to and not giving it a static address i created a dhcp pool inside docker and made nesseary changes to my pool for that net currently got mt fingers crossed. will update title after a few preliminary days of trial. thx guys

[skippy4hammypc]

Marked as Solved. Changed Docker to have DHCP server for Br4 (10Gbe), Configured the offending container to use Br0 with no static IP and grab an allocation from Docker itself. no issues after 18 hours of up time curiously I did still got one MACVLAN error in the syslog. hope that doesn't comeback to haunt me. 🤞

jarvlos-syslog-20210812-1738.zip