You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Dec 21, 2021. It is now read-only.
The idea here is to support running preparatory or maintenance tasks that need to execute as a command that finishes after a while.
Ideally this should also be done via systemd as some of these commands can run for quite a while (days sometimes) and we do not want them aborted, when the agent crashes or is restarted.
The job controller creates a pod with the label job-name=<name of the job>.
If the pod terminates successfully then its status contains phase: Succeeded and status/containerStatuses/state/terminated contains exitCode: 0 and reason: Completed.
If the pod terminates unsuccessfully then its status contains phase: Failed and status/containerStatuses/state/terminated contains exitCode: <exit code> and reason: Error.
The job does not delete the pod so that the result can be inspected.
(The job controller actually does a lot more but that is not relevant here.)
We have this ticket stackabletech/documentation#28 which talks about designing this feature.
That has been superseded by this one.
Is this a thing/change that needs to be documented? I believe so, right? We should document the Agent behavior.
I'll close the other issue in favor of this one.
The idea here is to support running preparatory or maintenance tasks that need to execute as a command that finishes after a while.
Ideally this should also be done via systemd as some of these commands can run for quite a while (days sometimes) and we do not want them aborted, when the agent crashes or is restarted.
This could be done via the following settings:
An example can be seen here.
I am unsure if RemainAfterExit is necessary, but this might be required to check the exit code of the command afterwards.
The text was updated successfully, but these errors were encountered: