This is a recurring issue, and I'm hoping that I find a resolution this time. I had the 'fixcommon problems' troubleshooting mode running for what that's worth. 

 

The symptoms are that I lose everything but ping. I can't ssh, telnet, or do web management. All my docker containers go unresponsive. It appears to have quit at 1:32AM. 

 

Any suggestions would be greatly appreciated. 

 

unraid-diagnostics-20170224-0102.zip

unraid-diagnostics-20170224-0132.zip

syslog.txt

OpenVPN wound up needing memory, none was available, so the system began to kill off processes. At the time of the first OOM, plex was using 62% of your RAM.

 

Plex wound up at one point utilizing 62% of the available RAM, and there were multiple killings of its processes (Invoked by various other programs).  What I think happened is that even after it was killed, the extreme amount of memory it was using wasn't reclaimed, and eventually another OOM occurred an this time a process more necessary to the operation of the system was killed off.

OK, thanks. I was having to rebuild Plex from scratch, and there's a bunch of media. If nothing else I could have run separate library additions. 

 

I thought part of the deal with Docker was to keep things discrete to keep them from interfering this way? 

It is.  Remember that it wasn't Plex that was invoking the out of memory process killings.  It was other processes that were attempting to kill Plex as there was nothing left for anything else at all.  You can limit Plex's resource usage (or any docker app for that matter) to any arbitrary value.  Check out the Docker FAQ (although a moderator inadvertently deleted the whole thing today -> might take a day to restore it), but if I recall correctly, edit the container, hit advanced in the top right, and in the extra parameters section enter in --memory=4G which will limit it to 4Gig of RAM instead of having carte blance on the whole system

 

Edited February 25, 20179 yr by Squid