We are trying out 4.12.3 and observe that once the cluster is stopped (shutdown) and started again it takes bit of time for MCO to get the lease (~5 mins) after API server is up,

$ oc version
Client Version: 4.12.3
Kustomize Version: v4.5.7

Every single time for us (CRC team) where we provision the openshift cluster and then use it for each run.

- Create a single node cluster and shut it down
- Start it again and as soon as apiserver is up edit/patch the pull secret
- Check the MCD logs and you will find the timestamp around how long it take to get the lease ownership (In our case it is ~5 mins) 

Takes around ~5 mins

shouldn't take more than 40-50 sec

Only workaround to make it faster is to remove old lease

oc delete lease machine-config-controller -n openshift-machine-config-operator