January 9, 201610 yr In the last week I have had two instances of unraid crashing (meaning the webinterface, shares and dockers remain functioning). Every time there are Defunct processes, as far as I can tell a defunct proces is a process that has ended but was not terminated properly. I am guessing that the fact that the processes are not correctly terminated is causing the webinterface to stall ("waiting" for a process to finish which it will never do).. Defunct processes cannot be KILL -9'd.. So it looks like I am stuck needing a reboot.. I have attached syslog but cannot find anything usefull in there.. syslog.zip
January 9, 201610 yr Author Trying to fix this without rebooting... From commandline I did: docker ps -a This gives me as list of running dockers including there ID (first collumn) Then I use docker stop <ID> to stop every running docker container. All dockers stop gracefully but for the SABNZBD container (needo version), this is taking a looooong time and showing the following error: kernel:unregister_netdevice: waiting for lo to become free. Usage count = 1 I will wait a bit to see if it solves it by itself, but this seems the culprit.. It feels like if I can get this to terminate (possible by stopping docker itself if it cannot solve itself that will cause the webgui to become responsive again.. With a bit of luck I will be able to restart the dockers again from the webgui..
January 9, 201610 yr Author Found the issue... The SABNZBD/NEEDO plugin will crash the webgui if I try to stop or restart it.. The restart of stop process will hang indefinately. If I kill -9 to process that is doing the restart then the webgui is back again..
January 9, 201610 yr Author Unfortunately the defunct process is still there.. The webgui however was available again... I tried to restart the server (thru the webgui), unfortunately that causes a gui crash again.. Powerdown from commandline shows the following error again: kernel:unregister_netdevice: waiting for lo to become free. Usage count = 1 p disk Shutting down does not complete.. stuck again...
January 9, 201610 yr Author Eventually got it responsive by shooting off some running processes by hand.. After that the powerdown script was able to run.. It gave a lot of errors but I think it stopped the array gracefully... System is now rebooting...
February 8, 201610 yr Mover is breaking for me sometimes: root@Tower:~# ps aux |grep rsync root 21647 0.0 0.0 11780 1448 ? D 03:40 0:00 rsync -i -dIWRpEAXogt --numeric-ids --inplace ./Video/TV/Transparent/Season 01/Transparent.S01E08.720p.WEB-DL-BATV.en.srt /mnt/user0/ root 21648 0.0 0.0 0 0 ? Z 03:40 0:00 [rsync] <defunct> root 24486 0.0 0.0 5104 1640 pts/4 S+ 10:24 0:00 grep rsync That subtitle file was a few hundred kb, and I have to hard reset my server, since I can't unmount the drive due to it being locked up by the failed rsync.
Archived
This topic is now archived and is closed to further replies.