Closed andrew-edgar closed 3 years ago
@andrew-edgar Did you test the new readiness probe? It always fails for me, most likely because the server only listens on the 127.0.0.1 address, but k8s will probe the actual pod IP address: https://github.com/cloudfoundry/capi-k8s-release/issues/91
So I think this PR can only be merged after we have bumped cf-deployment to something more recent.
I had tested the readiness probe and it did work for me. but maybe some of our other changes caused that to work. picking up 16.9.0 seems the right thing if that helps. excellent.
@jandubois yes we are now seeing in the older version the readiness check failing. I think we will get this fixed like you say when we bump to the latest cf-deployment. thanks
Add back the cc_deployment updater to the scheduler pod.
Description
Basically this reverted commit dfcc97d441efab31ec49a07b3cf87062baf7aeac. But also fixed the readiness check
Motivation and Context
This will allow the cc_deployment_updater to run and reenable the zero downtime functionality in CF.
How Has This Been Tested?
Deployed this in a full kubecf env and ran the CATs tests which validate this functionality
Screenshots (if appropriate):
Types of changes
Checklist: