coredns: Relax zonal topologySpreadConstraints

[hakman]

/cc @rifelpet @ameukam

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign justinsb for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[hakman]

/override pull-kops-e2e-aws-upgrade-k132-ko132-to-klatest-kolatest-many-addons

[k8s-ci-robot]

@hakman: Overrode contexts on behalf of hakman: pull-kops-e2e-aws-upgrade-k132-ko132-to-klatest-kolatest-many-addons

In response to this:

/override pull-kops-e2e-aws-upgrade-k132-ko132-to-klatest-kolatest-many-addons

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[jim-barber-he]

I just hit this on a test cluster where all its nodes are in a single availability zone, so only one of the 3 coredns pods would start.

Rather than removing the constraint, would it be better to keep it and just set whenUnsatisfiable: ScheduleAnyway?
That way where multiple zones are available it will still try to spread the pods across them, but not fail if it doesn't have that option?

[hakman]

/hold

[rifelpet]

The downside of ScheduleAnyway is that if there is room on an existing node in an undesired zone, k8s will schedule it there rather than keep it in Pending for your autoscaler to launch a new node in the desired zone.

Some users may find this acceptable while others may not.