mongo:latest crashes on restart

[vicary]

It runs well on the first download, but when I stop it never start again.

Here is the last lines of stdout.

2015-07-01T12:50:15.888+0000 W -        [initandlisten] Detected unclean shutdown - /data/db/mongod.lock is not empty.
2015-07-01T12:50:15.897+0000 I STORAGE  [initandlisten] ************** 
old lock file: /data/db/mongod.lock.  probably means unclean shutdown,
but there are no journal files to recover.
this is likely human error or filesystem corruption.
please make sure that your journal directory is mounted.
found 1 dbs.
see: http://dochub.mongodb.org/core/repair for more information
*************
2015-07-01T12:50:15.904+0000 I STORAGE  [initandlisten] exception in initAndListen: 12596 old lock file, terminating
2015-07-01T12:50:15.904+0000 I CONTROL  [initandlisten] dbexit:  rc: 100

Sorry I'm quite new to docker, while you guys are fixing this, can I like ssh in and delete the lock file myself? How can I do so?

[TeckniX]

@vicary you can do the following to attach to the container and do some clean-up:
$ docker-machine ssh dev
(will log you in to the actually VM)

docker@dev:~$ docker start -ai <container-id>

simply replace the container-id above with your monogodb container id obtained via docker ps -a
That should provide you with shell access to clean-up the data and then start the container.

[jmorganca]

Thanks for the issue, are volumes enabled?

[vicary]

@TeckniX Thanks!

@JeffDM Volumes are not mapped to Mac folders. Interestingly enough, it restarts normally in another Mac. Not sure but I will try again and update here.

[vicary]

Funny enough, I've got the steps to fix this:

1. Mount the /data/db volume to Mac
2. Remove it from setting
3. Docker will start by kitematic automatically
4. Mounted folder is empty so it can be started
5. docker exec mongo rm /data/db/mongod.lock

Then it can be restarted normally.

Since this is a Macbook Air I suspect it is related to the Mac sleep affecting VM state and ultimately killing the mongod process without proper shutdown.

[virgafox]

I'm experiencing the same issue. I use mongodb inside a container to keep my machine clean during development. If I start and stop it using CLI no problem, using Kitematic if I use the stop button I can't restart the container anymore with that error. I'm using Docker toolbox 1.8.1b. No shared volumes enabled, only binded port 27017.

[FrenchBen]

@virgafox Kitematic shouldn't do anything different than the CLI.
Did you create the container via the command line? If so, can you share the command arguments you used to create the container?

[virgafox]

@FrenchBen The container is always created via CLI, I couldn't find a way to bind ports creating via kitematic. The command is docker run --name mongodb -p 27017:27017 -d mongo. If I stop the container via the stop button in kitematic, I can't restart it anymore in any way, having the same error explained in this issue. If I stop it with docker stop mongodb I can restart it via cli with no problems.

[FrenchBen]

@virgafox Thanks for the break down of the issue, which I was able to replicate. I found the source of the #bug and will be pushing a create a patch/PR soon to get this fixed!

[virgafox]

@FrenchBen Thanks man. Maybe this is a little OT, but If I installed all via docker toolbox, how do I upgrade when the update is out? thanks again

[FrenchBen]

I believe you'll be able to download the latest toolbox and run the installer from it.

[virgafox]

@FrenchBen nice

[mchiang0610]

Seems like an issue with Official images. The team is notified.