Skip to content

Conversation

cbandy
Copy link
Owner

@cbandy cbandy commented Mar 8, 2022

Trying it out.

cbandy and others added 4 commits March 8, 2022 09:48
We added this field while implementing rolling updates and chose to
populate it with semantics similar to the appsv1.Deployment controller
which ignores terminating Pods. However, our controller does *not*
ignore terminating PostgreSQL Pods because they are still writing to
disk and still participating in HA quorum.

So, this field now reflects the number of observed Pods, regardless of
their state. This is the same approach taken by the appsv1.StatefulSet
controller. When this field is zero, we can be relatively certain that
all PostgreSQL processes have stopped and disks are idle.

Issue: [sc-13696]
See: https://issue.k8s.io/107920
See: e0a885c
The tests were written before the status fields existed. This also
reduces the scope of variables used only for failure messages.
These tests are flaky in CI and the current failure messages do not show
why. Add a testing package that can format, sort, and filter corev1.Events.
@cbandy cbandy changed the base branch from replica-counts to master March 8, 2022 22:49
@cbandy cbandy closed this Mar 8, 2022
@cbandy cbandy reopened this Mar 8, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant