#7069 (LF is down )

by Luc Saffre · in Hosting (system administration) · assigned to Luc Saffre

Summary

Children

Comments

1 day, 22 hours ago by Luc Saffre about #7069 (LF is down ) : Sharif helped me this morning to find the two last issues Missing /usr/lib/tmpfiles.d/daphne.conf The media of voga1r must be a link to the media of voga1e.
3 days, 4 hours ago by Luc Saffre about #7069 (LF is down ) : I'm still clueless after another session. systemctl status looks okay, healthcheck.sh says OK, but noi1r and voga1r report 404.
3 days, 16 hours ago by Luc Saffre about #7069 (LF is down ) : TODO: journalctl -u supervisor reports:
okt 22 13:56:10 liiv supervisord[567182]: ValueError: Could not create FastCGI socket tcp://127.0.0.1:8001: [Errno 98] Address already in use
3 days, 17 hours ago by Luc Saffre about #7069 (LF is down ) : I adapted the other demo sites on liiv. Now I'm running pull_and_prep_demos.sh (pm prep in voga1r was taking so long because I had forgotten to force periods.start_year...).
3 days, 18 hours ago by Luc Saffre about #7069 (LF is down ) : The daphne process is constantly stopping and restaring:
okt 22 11:59:26 liiv cosi1e[552753]: Started /usr/local/lino/shared/env/master/bin/daphne -u /run/daphne/cosi1e_daphne0.sock --fd 0 --access-log - --proxy-headers asgi:application (using lino_local.cosi1e.settings) --> PID 552753 okt 22 11:59:...
3 days, 21 hours ago by Luc Saffre in reply to Luc Saffre about #7069 (LF is down ) : I found one more error in the nginx config file. Still it is saying 504 Gateway Time-out. Maybe because the task runner is running... Because "journalctl -t cosi1e -" reports the following last line: okt 22 09:40:13 liiv cosi1e[512413]: Start task runner using <Logger lino (INFO)>...
4 days, 14 hours ago by Luc Saffre about #7069 (LF is down ) (1 reply) : I upgraded the Debian on the LF server, I changed the hostname to "liiv" because "lf" looked too much like "If". I now need to manually restore each Lino site. I'm still stuck with the first one, cosi1e, which theoretically should work but actually doesn't.
4 days, 19 hours ago by Luc Saffre about #7069 (LF is down ) : I will now upgrade the LF server. Remove gitlab-runner, docker and dovecot because I don't want them on the new system. Removed postfix because I prefer to reinstall it from scratch. It's an upgrade from Debian 12 to 13.
There's no real data on this server, only the config of the 4 demo sites. And t...
4 days, 22 hours ago by Luc Saffre about #7069 (LF is down ) : sudo tail -f /var/log/nginx/access.log shows that the server is being scanned. But also this looks kind of normal.
4 days, 22 hours ago by Luc Saffre about #7069 (LF is down ) : Most of the memory is used by uwsgi processes, 5 of which alone for cosi1e. And I don't see where this is configured. /etc/nginx/sites-enabled/cosi1e.conf looks normal.
4 days, 22 hours ago by Luc Saffre about #7069 (LF is down ) : Even after restarting all services (supervisor, nginx, mysql, gitlab-runner) the server is terribly slow. And I don't see why.
docker system df confirms that GL is installed but not used.
4 days, 23 hours ago by Luc Saffre about #7069 (LF is down ) : Seems that it is still not working correctly. Hm.
5 days, 12 hours ago by Luc Saffre about #7069 (LF is down ) : I disabled the voga1e demo site, to reduce the number of Lino sites on LF from 4 to 3.
5 days, 12 hours ago by Luc Saffre about #7069 (LF is down ) : LF is still using uwsgi instead of daphne because it runs an old Debian version. We didn't yet upgrade it because upgrading the mail server will probably cause much work. I consider dropping the lino-framework mailserver.
5 days, 16 hours ago by Luc Saffre about #7069 (LF is down ) : I started "nohup pull_and_prep_demos.sh" once more (the third time today).
5 days, 16 hours ago by Luc Saffre about #7069 (LF is down ) : It has been non-responding for hours now. I rebooted it using the OVH web interface. Now everything seems normal again.
5 days, 17 hours ago by Luc Saffre about #7069 (LF is down ) : Created new Ticket #7069 (LF is down ).