For the last 3 hours I have been watching the episodes of Planet Earth 2, streamed to my tv using Plex. At the same time another Plex video stream goes to my PC and a third Emby video stream goes to yet another PC. Meanwhile 10 container web GUIs are opened to serve requests.

 

And of course the unRAID GUI itself is open. No crashes, no hangups, all works perfect when clicking thru the GUI and obtaining all kinds of page information 

 

Can’t stop them all, one of them, a node docker for the Sonos api that I added wouldn’t stop. The others all have.

 

I’ve subsequently tried to restart the docker service with 

/etc/rc.d/rc.docker restart

but that’s not working, looks like I might be doing a forced reboot in a bit and then disabling that particular docker

Just to chime in... I have the same issues/symptoms as the other folks who have posted here.

7 minutes ago, bdillahu said:

Just to chime in... I have the same issues/symptoms as the other folks who have posted here.

 

Please check if you are able to stop your containers.

5 minutes ago, bonienl said:

 

Please check if you are able to stop your containers.

 

All but one of them.. "minio-unraid" won't stop...

 

I'm not docker expert, so not quite sure how to debug, but willing to try if you can tell me things to look for.

Try to read the details from that container

docker inspect minio-unraid

 

14 minutes ago, bonienl said:

Try to read the details from that container

docker inspect minio-unraid

 

 

Just hangs, doesn't respond...

So, I killed every process with "docker" that I could find and restarted rc.docker (cd /etc/rc.d; rc.docker start) and things seem to be working for now...

unRAID 6.5.1-rc2 is just released with some fixes. Please try this version.

 

Can you test again with unRAID-6.5.1-rc2.

 

This version has the correction for the TCP bug and addresses the issue when docker "hangs" on reading container information.

 

5 minutes ago, rix said:

Just as I posted it happened again.

 

Mar 24 11:46:26 ninja nginx: 2018/03/24 11:46:26 [error] 6302#6302: *15078 upstream timed out (110: Connection timed out) while reading response header from upstream, client: XX.XX.XX.XX, server: , request: "POST /webGui/include/DashboardApps.php HTTP/2.0", upstream: "fastcgi://unix:/var/run/php5-fpm.sock", host: "XXYYZZ.unraid.net", referrer: "https://XXYYZZ.unraid.net/Dashboard"

This time running the 6.5.1 RC2

 

If you wait long enough, does the list of Docker containers eventually appear or you keep a spinner on that page?

 

Edited March 24, 20188 yr by bonienl

No the php-fpm process is not the problem.

 

If you don't mind some 'hacking'.

In the file  /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php is the following statement at the near top.

return exec("timeout 20 /usr/bin/docker $cmd 2>/dev/null",$var);

Can you change the timeout value of 20 to 5.

After the change simply reload the Docker page in your browser.

Edited March 24, 20188 yr by bonienl

3 hours ago, rix said:

The SYN flooding one is new to me.

 

Harmless, this comes from Tautulli, I see this as well.

 

I have installed & configured pi-hole, but can not recreate the issue. Pi-hole is happily running (and blocking ads), I can start/stop it without issue. Docker page stays accessible all the time.

Thanks for testing with me.

 

With pi-hole stopped can you get  its information:

docker inspect pi-hole

 

Your container is declared "unhealthy", which is the only difference I see with my container.

"Health": {
                "Status": "unhealthy",

I am not sure how to get it "healthy", perhaps deleting the container+image and re-install may help?

 

May I ask everybody who had the Docker tab hanging to execute the below command (just copy + paste) and see if this succeeds.

docker inspect --format='{{.State.Running}}#{{range $p,$c := .HostConfig.PortBindings}}{{$p}}:{{(index $c 0).HostPort}}|{{end}}#{{range $p,$c := .Config.ExposedPorts}}{{$p}}|{{end}}#{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}#{{range $c := .HostConfig.Binds}}{{$c}}|{{end}}' $(docker ps -a --format='{{.Names}}')

Thanks

Just upgraded again to 6.5.0 to test this, but will have to downgrade before sunday evening because I am a week away and cannot restart the server.

Currently it looks to be running and your command just showed me all my containers on true except two, which had false.

 

 

1 minute ago, Uncledome said:

Just upgraded again to 6.5.0 to test this, but will have to downgrade before sunday evening because I am a week away and cannot restart the server.

Currently it looks to be running and your command just showed me all my containers on true except two, which had false.

 

 

 

Ok, thanks for testing. Please try this command again when you have the "failing" page situation.

 

Just now, bonienl said:

 

Ok, thanks for testing. Please try this command again when you have the "failing" page situation.

 

Will do and will let you know when and what happens.

I am having this issue as well, I am attempting to figure out what is causing the hanging, only recent system change besides upgrading to 6.5.0 was adding another nic and changing to active-backup. Same symptoms as everyone else.

 

 

4 hours ago, bonienl said:

docker inspect --format='{{.State.Running}}#{{range $p,$c := .HostConfig.PortBindings}}{{$p}}:{{(index $c 0).HostPort}}|{{end}}#{{range $p,$c := .Config.ExposedPorts}}{{$p}}|{{end}}#{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}#{{range $c := .HostConfig.Binds}}{{$c}}|{{end}}' $(docker ps -a --format='{{.Names}}')

 

This command and docker stats hang for me as well. Nothing unusually memory or cpu usage wise either. I will report back if I figure anything out.

 

 

After running docker stats --no-stream on each container separately it only would hang on pi-hole. docker stop and docker kill would not work on the pi-hole container either so I had to kill the process. Once that was done everything was responsive again. I will try to figure out what the issue is with pi-hole tomorrow if I have time as I would like to continue running it. Thanks to bonienl and everyone else in this thread for helping to point me in the right direction.

 

So not sure if this will work long term but I will report back..for 2 days pi-hole has been running with no issues. I changed br0 to bond0 and that was it. I thought it might have something to do with networking after reading about a more general docker issue with hanging because of network settings in the docker. Bond0 works for me as I am running with two network cards bonded as active-backup. This was for the interface variable in pi-hole. Good luck to anyone else having this issue but everything that has been outlined will at least help you identify the docker causing your issue!

Edited March 26, 20188 yr by cschanot

Just an update from me, with my troublesome docker turned off everything is working fine, it's not Pi-Hole for me but rather this : https://github.com/chrisns/docker-node-sonos-http-api perhaps there are similarilties, although not obviously to me!

I've not had a chance to update to RC2 yet.

For me after several rounds with reboots it seems to have fixed itself. The problem hasn't come back in the last 5 days.

FWIW I have the same timeout issue starting with 6.4.1. Didn't have it on 6.3.x and was hoping 6.5.0 would fix problem.