Closed amotl closed 4 years ago
We added the following bits to our systemd configuration in order to mitigate this issue.
cat /etc/systemd/system/redis.service.d/restart.conf
:
[Service]
# https://jonarcher.info/2015/08/ensure-systemd-services-restart-on-failure/
# Please run "systemctl daemon-reload" after making changes to this file.
Restart=always
RestartSec=3
Please run systemctl daemon-reload
after adding this file or making changes to it.
While we recognize this is just a lousy workaround, it helped us tremendously for just keeping the whole system up and running. Saying this, we should keep in mind that we someday should investigate the root cause why Luftdatenpumpe is regularly crashing Redis.
We also updated the documentation and added the relevant information at [1].
[1] https://github.com/panodata/luftdatenpumpe/blob/master/doc/setup/redis.rst#running-in-production
We and others noticed that while running Luftdatenpumpe in production, the Redis server frequently fails. On a subsequent invocation,
luftdatenpumpe
will fail likewise.