rudder service and thus cf-execd is never restarted / ensure running
In 4.1, we removed the check that cf-execd was running at least/at most a number of time because it used to cause problems with containers.
But we now don't check anymore that the rudder service is running at all, so it means that if cf-execd is stopped (for example, because the rudder service is not started at boot), it is never restarted.
Rudder agent will still run because of /etc/cron.d/rudder-agent, but rudder agent health will report problems (no outputs stored in CFEngine output directory), and probes checking for cf-execd status will be in error.
Finally, we need to check that rudder (and not cf-execd) service is running to well behave with systemd.
A workaround can be to create a technique "ensure service running" for rudder service and deploy it everywhere.
#2 Updated by François ARMAND 7 months ago
- Assignee set to Nicolas CHARLES
- Effort required set to Small
- Priority changed from 77 to 80
This is even worsen in the case where agent schedule is NOT 5 minutes, because the cron actually runs every 5 minutes. Before that bug, it would means that we add at most one run too many. Not anymore.
#7 Updated by Nicolas CHARLES 5 months ago
- Status changed from In progress to Pending technical review
- Assignee changed from Nicolas CHARLES to Alexis MOUSSET
- Pull Request set to https://github.com/Normation/rudder-agent/pull/134
#8 Updated by Nicolas CHARLES 5 months ago
- Status changed from Pending technical review to Pending release
Applied in changeset rudder-agent|6fc2de2971da49d161617166c232d3dd07c7dc4b.
#9 Updated by Vincent MEMBRÉ 3 months ago
- Status changed from Pending release to Released
- Priority changed from 79 to 96
This bug has been fixed in Rudder 4.1.9 and 4.2.3 which were released today.
- 4.1.9: Announce Changelog
- 4.2.3: Announce Changelog
- Download: https://www.rudder-project.org/site/get-rudder/downloads/