Warn if notify_push daemon is not running ?

[Mer0me]

For some reasons, the notify_push daemon wasn't running anymore on our instance. That didn't throw any error, and seems not to be logged anywhere. Consequences : talk notifications were not working at all, or very very very slowly (several minutes).

systemctl enable --now notify_push solves the problem instantly. But it takes me hours to figure it out.

Maybe the client push app should warn somewhere if the daemon is not running ?

[EnderArchery]

For me service sometimes doesn't start as 127.0.0.1 is not permitted to connect to the sql-server. I suggest adding After = mysql.service to the unit using the setup-assistant, if the nextcloud instance has a database configured that's running on the localhost.

There might be better ways to do it though. But I think it's definitely a problem worth looking into... at some point?

[reos-rcrozier]

I had a similar issue, solved by using a bash script something like

is_active=$(systemctl is-active notify_push.service)


if [ "${is_active}" = "active" ] ; then
   echo "service notify_push is active."
 else
   echo "systemctl indicates that the notify_push service is not active." | mutt -s "Services Alert for notify_push" sysadmin@<redacted>.tld
 fi

Then I put this in a cron job and it emails me (using mutt) if it's not running.

[joshtrichards]

Related (along with some proposed solutions): #194