Hi,
I’m testing the upgrade from 6.0.4 to 6.2.9 (Professional server).
Using fastcgi everything work good, but when I switch to WSGI mode seahub.sh fails to start if thrown from the root account.
Seafile is running on an Ubuntu 14.04 without memcached.
I had to change the init script as follows, in order to make it works:
sudo -i -u ${user} ${script_path}/seahub.sh start >> ${seahub_init_log}
Adding the -i
option to sudo
I could workaround the issue, but I’d rather fix it.
This is the scenario:
-
root~# /etc/init.d/seafile start
- seafile.sh run and seahub.sh fails to start -
root~# sudo -u user /home/user/seafile/seafile-server-latest/seahub.sh start
- it fails after thepgrep -f seahub.wsgi:application
- Rebooting the server gunicorn is launched correctly, if I restart the service it fails
- user~# /home/user/seafile/seafile-server-latest/seahub.sh start - it works
- Adding the
-i
option atsudo
it works.
What I did so far that didn’t work:
- I checked for permissions issues on the seafile folder, running
chown -cR
- I debugged gunicorn using --log-level debug, but I can’t find anything relevant
- I debugged gunicorn using the
pdb -m
option at the python execution line in seahub.sh script - I deleted cache and logs in the runtime folder
- I checked in the forum for similar issues
Any ideas?
Thanks
P.S. the package with the “_Ubuntu.tar.gz” suffix when it should be used? I tried both, same result.