An interrupted service restart may leave services in the exited state
#2257
Comments
pipex
added a commit
that referenced
this issue
May 8, 2024
When uninstalling a container for a restart from the API on when moving between releases, this prefers to force remove the container rather than stop then remove. This has the same effect in the end, but force remove is an atomic call, meaning if the supervisor is killed between the stop and remove, the engine will proceed with the operation and the supervisor will know to recover from the target state after the restart. Change-type: patch Relates-to: #2257
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
A service restart via the supervisor API endpoints
performs the following actions
Because of the changes in #1926 if the supervisor gets stopped between steps 1 and 2, on restart, the service won't get restarted by the supervisor when applying the target state, and it won't be started by the engine since a
stoprequest already took placeThis might be solved by combining steps 1 and 2 by doing a remove with
force: truewhen killing a containerThe text was updated successfully, but these errors were encountered: