Continued issues with stopping dockers (6.5.2)

[fiore00713]

I'm still trying to track down the source of some issues I'm experiencing. These seem to have started after updating to 6.5.0 and are continued into 6.5.2

I'm having issues with dockers stopping, it is not affecting all dockers but some of my more 'important' ones.

Specifically, I am seeing issues with Plex and Nextcloud. When the dockers are attempting to stop I can consistently seeing in the logs:

 

[s6-finish] syncing disks.

 

I am also seeing some questionable stuff in the system log

 

Jun 27 20:19:09 unRAID nginx: 2018/06/27 20:19:09 [error] 15648#15648: *899976 upstream timed out (110: Connection timed out) while reading upstream, client: 192.168.1.216, server: , request: "POST /Docker/UpdateContainer?xmlTemplate=edit:/boot/config/plugins/dockerMan/templates-user/my-PlexMediaServer.xml HTTP/2.0", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "920a3f91f9c2670c85e3fe9f0fc5f37d7b95ac87.unraid.net", referrer: "https://920a3f91f9c2670c85e3fe9f0fc5f37d7b95ac87.unraid.net/Docker/UpdateContainer?xmlTemplate=edit:/boot/config/plugins/dockerMan/templates-user/my-PlexMediaServer.xml"

 

Linked is a previous thread that I had started but issues were not resolved then.

I have since updated to 6.5.2, I have not yet had time to update to 6.5.3 but can make time to do so if deemed necessary.

I have re-created my docker image, hoping it might be something wrong with that but issues still presents.

 

The biggest thing I seem to keep getting hung up on is the [s6-finish] syncing disks portion. I've read in some other threads that this was talking about the items in ram being written to the array. Wondering if there's anyone that could elaborate further on this? Is this items in ram from the specific docker or system wide? My unRAID server has 128gb of memory installed, of which about 40gb is allocated to various VMs that are running, I also am using ram transcoding within PLEX and the memory being used by the other dockers and processes, etc.. We could be talking about 50+GB trying to be written to the array. But even still, I cant imagine it would that that long for 50gb to be written. I have had instances where the docker hangs during stopping for 10-20 minutes (this is pretty rare, honestly) but it will generally hang for HOURS and eventually come back around and start acting normal.

 

Again, this is only happening with SOME of my dockers. I have 18-20 dockers that are always running in some capacity. The majority of which are set to auto update nightly and I don't have issues with. So I'm trying to figure out what about some of these dockers are making things grumpy.

 

Attached are logs. Currently I'm updating a setting in PLEX so the issue is happening NOW as of the time of the log generation

 

 

 

unraid-diagnostics-20180627-2039.zip