Consider caching the zone information for a datastore in vSphere

[subramanian-neelakantan]

What happened:
When creating a PV with the vSphere Cloud Provider (VCP), the volume plugin invokes vCenter Tag APIs to figure out the zone information for the newly created Volume. This computation happens each time a Volume is created on a datastore. This is a moderately heavy computation with multiple Tag API calls to the vCenter. This is a change proposed in PR 72687 and PR 72731 that fixes the zone usage of a Volume in vSphere.

What you expected to happen:
Since the zone information does not change often, it would be good for VCP to cache this. This cache needs to be built such that it can be initialized and used by kube-controller as well as the api-server during admission control.

However, note that a change in the Tags applied in vCenter, or a change in vCenter topology, could invalidate this cache. This could demand a restart of these binaries to force rebuild this cache. So this change needs to be evaluated carefully.

Also, need to evaluate how much of a performance damage do the above PRs cause while creating a Volume. This issue should be based on whether the regression hurts, and whether there is a meaningful way to cache this information without stale entry invalidation.

How to reproduce it (as minimally and precisely as possible):
Create a PV with VCP configured in a k8s cluster. Note from the logs that both the kube-controller as well as api-server make several calls to VC in order to figure out the zone information for a datastore.

Environment:

• Kubernetes version (use kubectl version): any
• Cloud provider or hardware configuration: vsphere

[subramanian-neelakantan]

/sig vmware

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

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/test-infra repository.