Error: Cannot open an HTTP server: socket.error reported AF_UNIX path too long
I have installed galaxy in Ubuntu 20.04 as guided in https://galaxyproject.org/admin/get-galaxy/ I have updated an existing one by command:
$ git fetch origin && git checkout release_22.01 && git pull --ff-only origin release_22.01
But after “sh run.sh”, there comes some error as below. Please help with this. Thank in advance.
Executing: galaxy
2022-05-28 18:21:59,622 WARN No file matches via include "/home/ubuntu/Desktop/NV042022/Data_infras/GITHUB-ANSIBLE-REFERENCE/galaxy/database/gravity/supervisor/supervisord.conf.d/*.conf"
2022-05-28 18:21:59,622 INFO Included extra file "/home/ubuntu/Desktop/NV042022/Data_infras/GITHUB-ANSIBLE-REFERENCE/galaxy/database/gravity/supervisor/supervisord.conf.d/_default_.d/galaxy_celery-beat_celery-beat.conf" during parsing
2022-05-28 18:21:59,622 INFO Included extra file "/home/ubuntu/Desktop/NV042022/Data_infras/GITHUB-ANSIBLE-REFERENCE/galaxy/database/gravity/supervisor/supervisord.conf.d/_default_.d/galaxy_celery_celery.conf" during parsing
2022-05-28 18:21:59,622 INFO Included extra file "/home/ubuntu/Desktop/NV042022/Data_infras/GITHUB-ANSIBLE-REFERENCE/galaxy/database/gravity/supervisor/supervisord.conf.d/_default_.d/galaxy_gunicorn_gunicorn.conf" during parsing
Error: Cannot open an HTTP server: socket.error reported AF_UNIX path too long
For help, use /home/ubuntu/Desktop/NV042022/Data_infras/GITHUB-ANSIBLE-REFERENCE/galaxy/.venv/bin/supervisord
The path to the socket that we're using got too long in your setup. You can choose another path by setting GRAVITY_STATE_DIR=/home/ubuntu/some_shorter_path ./run.sh.
https://mail.openvswitch.org/pipermail/ovs-dev/2013-January/024528.html
UNIX socket paths have a limit in terms of length. On Linux this is 103 characters.
wow that is a short limit. https://unix.stackexchange.com/questions/367008/why-is-socket-path-length-limited-to-a-hundred-chars
We should provide a link to instructions here with the error message.
Totally forgot about this, it is likely to be a semi-frequent occurrence. I wonder if we should drop the setting of $GRAVITY_STATE_DIR in the venv activate script and just use the default of ~/.config/galaxy-gravity, plus a shortened hash of the galaxy config file path to separate instance configs. We can default the logs to somewhere in the Galaxy dir.
Just hit this on macos 12.6, would be nice to see a more actionable error given the workaround is pretty easy.