Details
-
Suggestion
-
Resolution: Unresolved
-
None
-
None
-
1
-
1
-
Description
Currently, we're pushing as much of new software deployments to RHEL/CentOS 7. This morning when I came in, I found that our Confluence server's java process had exited. I was able to login and `systemctl restart confluence` to get it back online, but it would be helpful if Confluence's SysV-style init scripts were supplemented with systemd-aware service-management components. If Confluence were systemd-aware, then crashing of the main Java process likely could be made to auto-restart.
Yeah, we can construct our Confluence deployments to re-deploy on service-loss, but letting systemd handle simple cases such as this morning would be quicker than a re-deployment and make the question of clustering about scale rather than about availability.