Service not reliably active after install
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
pcs (Ubuntu) |
Expired
|
Undecided
|
Unassigned |
Bug Description
Hi,
the daily metric runs, among other things, install pcs.
In there I've had two days where in mantic the service did not come up.
Affected:
- 12th+13th of July
- Good before & after
What I see in the bad case is this:
08:05:59 Created symlink /etc/systemd/
08:05:59 Created symlink /etc/systemd/
08:06:02 Could not execute systemctl: at /usr/bin/
Full logs:
good: https:/
bad: https:/
That seems to be the start call at the end of postinst
deb-systemd-
in
/var/lib/
Due to that
OTOH, a first thought might be "why isn't is failing the install then?"
Well that is easy, this has an || true (see above).
So is this expected?
Why?
Would then a retry mechanism make sense?
I've created (https:/
Please have a look if you think this something we can improve with what we have so far.
If not close it as incomplete until it re-occurs or someone else chimes in having more data.
Thanks for reporting this bug Christian.
As you may have anticipated, I was not able to reproduce this issue locally. That command in the postinst script you mentioned is autogenerated by dh_installsystemd so nothing manually crafted by the Debian maintainer. I think we shouldn't manually change this behavior right now, if you find any occurrence of that again we can revisit and re-asses the situation. Does that sound good to you?