There's a different root cause issue here - a new bug 1879598 has been opened to track it.
This fix was for the case where there was a race execing into the workload pod as its init containers were running.
The new issue appears related to pod termination when a rolling update is done by the k8s deployment controller.
It's been hard to reproduce - any chance of getting access to the mattermost charm so we can try that to see if we can get things to misbehave?
There's a different root cause issue here - a new bug 1879598 has been opened to track it.
This fix was for the case where there was a race execing into the workload pod as its init containers were running.
The new issue appears related to pod termination when a rolling update is done by the k8s deployment controller.
It's been hard to reproduce - any chance of getting access to the mattermost charm so we can try that to see if we can get things to misbehave?