No Docker Tab

I reinstalled my UNRAID v6.0.0 cleanly today as I was having a few issues with Docker containers, I thought maybe it was because of upgrading since the beginning of the v6 betas.

So to do this as cleanly as possible I formatted my USB drive for my clean v6 installation.

I moved my VMs and docker.img from the cache drive to another disk so I could format this too (I wanted to be sure it was using the latest btrfs file system).

 

Once started and I reassigned my drives to the correct slots, I started the array and copied over the docker.img file and enabled docker in the settings with the same config as before (size and path).

 

I installed a few plugins that have been really useful in the past (nerd tools, community applications, Dynamix system stats and docker search)

 

It was then I noticed (it may have been before installing these however) that I have no Docker tab, I tried trying the url http://myip/Docker but htis only shows a blank page below the Limetech header.

 

Here are the logs from when I tried starting docker,

 

Jun 24 12:09:57 msstorage php: /usr/local/emhttp/plugins/dynamix.docker.manager/event/started;/usr/local/emhttp/plugins/dynamix.docker.manager/dockerupdate.php

Jun 24 12:09:57 msstorage php: Starting Docker...

Jun 24 12:09:57 msstorage kernel: BTRFS info (device loop0): disk space caching is enabled

Jun 24 12:09:57 msstorage kernel: BTRFS: has skinny extents

Jun 24 12:09:57 msstorage kernel: BTRFS (device loop0): bad fsid on block 20987904

Jun 24 12:09:57 msstorage kernel: BTRFS (device loop0): bad fsid on block 20987904

Jun 24 12:09:57 msstorage kernel: BTRFS: failed to read chunk root on loop0

Jun 24 12:09:57 msstorage kernel: BTRFS: open_ctree failed

Jun 24 12:09:57 msstorage logger: Not starting Docker: mount error

Jun 24 12:09:58 msstorage php: Updating templates... Updating info...

Jun 24 12:09:58 msstorage php: Warning: stream_socket_client(): unable to connect to unix:///var/run/docker.sock (No such file or directory) in /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php on line 505

Jun 24 12:09:58 msstorage php: Couldn't create socket: [2] No such file or directory Done.

 

Do you think this points to an error with my docker.img or my cache disk? The disk is old but UNRAID doesn't report any errors here.

 

Any ideas on how I can get my docker tab back?

 

Do you all run docker containers without any problems? The main issue that I had before was checking for containers updates then watching as most of the containers dies on me and would not restart until I restarted the UNRAID server.

Jun 24 12:09:57 msstorage logger: Not starting Docker: mount error

 

that leads me to believe the path in the docker settings is wrong. disable docker, repoint it to the img using the menu that pops up when selecting the input box (make sure it is 100% cache only (if using a share)) then try to re-enable it. also I would check for lower and uppercase issues int he path (i made this mistake once) and compare it to the folder and file names using MC or winscp.

The path is correct (see attachments)

 

I appreciate your help.

Stop docker.

Rename your docker.img to docker-back.img

Start docker to create a new docker.img file.

last thing I can think of is the perms.

 

added a snap of the permissions on mine from MC

 

If that doesn't work, Looks like your starting over and building from scratch like mentioned above.

 

ok, it seems like something went wrong during the copy.

 

I deleted the .img that I copied this morning and then recreated an empty docker.img of the same size, stopped docker, renamed this and copied the docker.img I had saved and it's all working.

 

Thanks for your help guys.

Spoke too soon....

Docker would start ok, I saw the list of containers although without any icons and the only option if I clicked on them were "Start", no edit or logs etc.

 

I'll install them all again using a clean docker.img

 

I had quite a lot of bad experiences using docker with UNRAID, what is the general opinion guys?

Always working without issues?

Sometimes?

Often problematic?

 

Thanks for your thoughts.

Remember, using a clean docker.img file does not mean you loose all your settings!

 

The settings are stored in the my-templates files on the flash drive somewhere. Recreating the docker.img file means you'll have to re-download all the images again.

I have never tried to pull over an existing docker. When I swapped in an SSD for my cache from the HDD. I just remade everything. I like clean starts personally.

 

I did run into the issue of naming the folder "Docker" vs "docker" and couldn't figure out why it wouldn't work. Then realized windows isn't case sensitive like *nix. 2 hours of life wasted.

Thanks for the advise guys.

I knew that using a clean docker.img file didn't lose my settings as I used the same config paths so I deleted the old docker.img and started cleanly this way.

 

I reinstalled my containers and all worked well for a day.

However when I tried to update the configuration to my beets container (which also ran ok for a day) I then "lost" the container, it's disappeared from the list. I thought this may have been the beets app so I tried to modify my "owncloud" container and tris has also disappeared from the web view in Docker.

 

So I tried to reinstall and I have the following error message,

 

Warning: file_put_contents(/var/lib/docker/unraid-update-status.json): failed to open stream: Read-only file system in /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php on line 297 Warning: file_put_contents(/var/lib/docker/unraid-update-status.json): failed to open stream: Read-only file system in /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php on line 453

 

Can anyone point me in the right direction please?

I'm loving the idea of docker containers and feels this brings so much to the already excellent UNRAID, however I'm losing faith :-(

Since my last post I checked if the file /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php was read only and it isn't.

 

I haven't had time to restart my server yet (this was the only solution to get the containers running again) but another container crashed (mineos).

 

The only container that stays working is Logitech Squeezebox, which was the case before when I had these problems.

Tools - Diagnostics. Post result.

Do you want every diagnostic file?

Or only docker and the log folder?

 

 

Do you want every diagnostic file?

Or only docker and the log folder?

 

Personally I'd post the lot.

There we go.

msstorage-diagnostics-20150628-2245.zip

I had the same sort of problem, after rebooting from 6.0.1 upgrade my dockers (BTSync and PlexMedia Server) would not start, when trying to start I would get message below. I rebooted again hoping to restore dockers only to then have them disappear from the webpage, I was unable to restore docker tab until i deleted the docker.img file and was then able to to restore my docker files to the webpage. also attached is a photo of a BTRFS error message that displayed during reboot (restored to 6.0 to see if 6.0,1 was causing problem which it wasn't, same thing happened with 6.0) all is well after re creating docker.img file but it did create a few anxious hours.

 

(Warning: file_put_contents(/var/lib/docker/unraid-update-status.json): failed to open stream: Read-only file system in /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php on line 297 Warning: file_put_contents(/var/lib/docker/unraid-update-status.json): failed to open stream: Read-only file system in /usr/local/emhttp/plugins/dynamix.docker.manager/dockerClient.php on line 453)

Slightly reassuring to know it's not only me but sorry you're having these problems too.

 

It'll be interesting if, after some time of your docker containers running that you try to modify a setting on one of them, that was generally what crashed the rest of my dockers and then requiring a restart of UNRAID and a recreation of the docker.img.

 

Any others with similar problems?

Appears to be a one time failure, switched BTSync to the limetech sync docker (to verify changes to dockers wasnt cuase of problem) and rebooted without problems. luckily recreating dockers is easy with the my-dockername option when creating dockers.

Glad it was a one off for you 

I also have weird issues with Docker. Overnight the containers crashes and I have to recreate the containers again. I upgraded form RC to 6.0.1. I have quickly run a diagnostic. See the link. I have to go to work so no time to investigate further.

 

https://dl.dropboxusercontent.com/u/4327743/tower-diagnostics-20150701-0602.zip

I also have weird issues with Docker. Overnight the containers crashes and I have to recreate the containers again. I upgraded form RC to 6.0.1. I have quickly run a diagnostic. See the link. I have to go to work so no time to investigate further.

 

https://dl.dropboxusercontent.com/u/4327743/tower-diagnostics-20150701-0602.zip

Your appdata share has default settings and mover is moving it to the array

Jul  1 03:40:01 Tower logger: mover started
Jul  1 03:40:01 Tower logger: moving "appdata"
Jul  1 03:40:01 Tower logger: ./appdata/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports
Jul  1 03:40:01 Tower logger: .d..t...... ./

You must set your appdata share to Use cache disk: Only.

 

Why didn't you have this problem with the RC? Did you do a clean install or something?

I also have weird issues with Docker. Overnight the containers crashes and I have to recreate the containers again. I upgraded form RC to 6.0.1. I have quickly run a diagnostic. See the link. I have to go to work so no time to investigate further.

 

https://dl.dropboxusercontent.com/u/4327743/tower-diagnostics-20150701-0602.zip

Your appdata share has default settings and mover is moving it to the array

Jul  1 03:40:01 Tower logger: mover started
Jul  1 03:40:01 Tower logger: moving "appdata"
Jul  1 03:40:01 Tower logger: ./appdata/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports
Jul  1 03:40:01 Tower logger: .d..t...... ./

You must set your appdata share to Use cache disk: Only.

 

Why didn't you have this problem with the RC? Did you do a clean install or something?

 

This really begs the question for me. Should dockers manager just create it's own default cache only share during initialization? Advanced users could use their own settings.

I also have weird issues with Docker. Overnight the containers crashes and I have to recreate the containers again. I upgraded form RC to 6.0.1. I have quickly run a diagnostic. See the link. I have to go to work so no time to investigate further.

 

https://dl.dropboxusercontent.com/u/4327743/tower-diagnostics-20150701-0602.zip

Your appdata share has default settings and mover is moving it to the array

Jul  1 03:40:01 Tower logger: mover started
Jul  1 03:40:01 Tower logger: moving "appdata"
Jul  1 03:40:01 Tower logger: ./appdata/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports
Jul  1 03:40:01 Tower logger: .d..t...... ./

You must set your appdata share to Use cache disk: Only.

 

Why didn't you have this problem with the RC? Did you do a clean install or something?

 

This really begs the question for me. Should dockers manager just create it's own default cache only share during initialization? Advanced users could use their own settings.

Sounds like you are calling for a separate cache-only share for each docker. That seems excessive and would possibly compound the problem if there was some problem where the .cfg for those shares wasn't accessible, such as might happen with a clean install.

 

There are some discussions about these issues ongoing here.

Nah, just the appdata share since it seems to be the template default.

Nah, just the appdata share since it seems to be the template default.

Squid is already on it with the latest version of Community Applications plugin.