Fix for deploying Arktos cluster with Mizar as default CNI plugin

click2cloud-team1 commented 3 years ago

What type of PR is this? Fixes

What this PR does / why we need it:

Using PR #1114, after Arktos cluster gets running, it takes a while to start the mizar-daemon, and once it gets running, we need to delete the mizar-operator pod for the first time. After CNI deployment, VPC and subnet are stuck in init state. Once the mizar-operator pod is recreated, all the network resources gets in provisioned state.
Once we redeploy the cluster, all the network resources gets provisioned.
This PR will resolve the above issue to avoid deletion of the mizar-operator pod for the first time.

vinaykul commented 3 years ago

We need to root-cause and understand why we have this issue in Arktos and not in upstream k8s. /hold

vinaykul commented 3 years ago

This change is not required.

CentaurusInfra / mizar