-
Notifications
You must be signed in to change notification settings - Fork 48
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Stop watching leases for custodian controller #688
Stop watching leases for custodian controller #688
Conversation
eb2522d
to
747be03
Compare
@unmarshall, @seshachalam-yv You have pull request review open invite, please check |
b1958b9
to
f225e10
Compare
/test pull-etcd-druid-e2e-kind-alpha-features |
@unmarshall thanks for your comments. I have addressed them. PTAL |
/retest |
/test pull-etcd-druid-e2e-kind |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
/lgtm
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @shreyas-s-rao Nice Fix
Saving a lot KAPI calls 🚀
/lgtm
How to categorize this PR?
/area quality
/kind bug
What this PR does / why we need it:
Custodian controller no longer watches lease objects, but rather does an initial reconciliation on any etcd object that has finished reconciliation, and requeues the event for after
SyncPeriod
interval, as was being done previously as well.This PR ensures that custodian controller does not update the etcd status too frequently, thus saving hundreds of calls to kube-apiserver every second per seed cluster, not to mention the reduction in number of failed etcd status update calls from etcd controller to update
observedGeneration
andlastError
information.Additionally, the predicate
EtcdReconciliationFinished
has been modified to check equality ofstatus.ObservedGeneration
andmetadata.Generation
, rather thanstatus.Ready
. Reason for this is that reconciliation is technically finished only when etcd reconciler marks the observedGeneration with the generation it "observes", and has nothing to do with whether the etcd cluster is ready or not.Which issue(s) this PR fixes:
Fixes #683
Special notes for your reviewer:
/invite @unmarshall @seshachalam-yv
/assign @unmarshall @seshachalam-yv
Release note: