Hello Mathieu, Thibault,
I think there are two problems here:
- We need to call "systemctl daemon-reload" on
shinken-common.postinst
=> This is needed for systemd to know about the new init file
Yes, I confirm that it solve the "first start" issue.
- We probably need to mask the shinken service, as this is a
bad-designed initscript which start all daemons from the same cgroup.
See #740942 for a similar problem
I agree and I think there is an other side effects to this.
For example when I execute a restart of all individuals services with
`services shinken-* restart` all daemons successfully restart and seem
to be alive but shinken "checks" seems to be stopped and never happen
anymore.
I solve this issue by restarting again only the poller (wait some time)
and then the arbiter. After that shinken checks reworks correctly.
So I think there is also something bad about the order/concurrency when
all daemons start.
Maybe this could help you solving this issue.
Best regards,
Thomas Pierson
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
