Dockerlaunch socket stuck after docker dies

[panch13114]

The dockerlaunch socket is recreated as a root-owned volume on docker restart if the docker daemon is unexpectedly killed

[philtweir]

In the glossia-server-side directory, try the following:

export COMPOSE_API_VERSION=$(docker version | grep 'Server API' | awk '{ print $NF }')
docker-compose -f docker-compose.local.yml rm

then go for ./start-local.sh

[panch13114]

I get the following error, after executing the above commands.

For the first command:
I get permission denied error
For the second command:
To include them, use docker-compose rm --all.
This will be the default behavior in the next version of Compose.

ERROR: Couldn't connect to Docker daemon at http+docker://localunixsocket - is it running?

If it's at a non-standard location, specify the URL with the DOCKER_HOST environment variable.

[philtweir]

Oh, sorry, you need to be root or using sudo. If you want, you could do it:

sudo docker version | grep 'Server API' | awk '{ print $NF }'

Should return 1.18 or 1.21, or similar - put that in literally at [VER] below.

sudo COMPOSE_API_VERSION=[VER] docker-compose -f docker-compose.local.yml rm --all

[panch13114]

I tried
sudo COMPOSE_API_VERSION=1.18 docker-compose -f docker-compose.local.yml rm --all
It removed most of glossia bar and run etc.
I started the docker service again and tried to launch dockerlaunchd.
Get the same error

[philtweir]

If you now stop docker (sudo service docker stop), then rm -rf /var/run/dockerlaunch, then start docker (sudo service docker start) does that directory appear again owned by root?

[panch13114]

Thank you. It worked

[philtweir]

OK, great - I'll leave this issue open, as we need to find a way of avoiding this in the first place...