High memory usage, all dockers freeze. Seems nginx related.

[Tuftuf]

Since upgrading to 6.5.0 I've had issues with all my dockers stopping working, in other words plex stops working along with everything else.

 

I saw that nginx was using a large portion of it and restarted it. Memory usage goes from 99% to 66% usage and all my dockers start to recover by themselves but I still can't see any dockers or VM's on the dashboard.

 

I've changed disk cache vm.dirty_background_ratio to 4 and vm.dirty_ratio to 8.

0327 - diagnostics was taken after restarting ngnix, and changing 

 

Since making the change to cache settings i've not had another docker crash, but again my server is sitting at high memory usage when last night it was around 70%
0329 - diagnostics was taken while memory is sitting at 96%.

 

Restarting nginx will reduce the memory usage to normal.

 

Any ideas? Anything I'm missing?

 

tower-diagnostics-20180327-1721.zip

tower-diagnostics-20180329-1234.zip

[John_M]

How are you measuring memory usage?

 

[Tuftuf]

 Before restarting ngnix.

 

After restarting it, nginx is now displayed on page 3 using 79Mb.

 

 

 

I admit my memory usage is generally high, but prior to the upgrade to 6.5.0 I've not seen my dockers all grind to a halt. 

Internet handled through a VM

DNS handled through a docker. So this isn't something I've run into regularly or I would have noticed.

 

VM continue to work when this issue occurs

Dockers all grind to a halt.

 

Restarting nginx free's up memory and everything springs back into action, but I'm trying to understand if nginx is problem or something else.

Edited March 29, 2018 by Tuftuf