Coredns fails to start if Corefile migration fails

[anttitapio]

What happened:
Coredns failed to start in 1.17 deployment when it tried to retain a backed up Corefile with error message: Caddyfile via flag: open /etc/coredns/Corefile: no such file or directory

What you expected to happen:
Coredns starts normally with the backed-up Corefile

How to reproduce it (as minimally and precisely as possible):
Deploy a k8s 1.17 cluster and trigger coredns Corefile migration and make it fail. In our case, this was triggered in a three master -setup with 'old' method of running 'kubeadm init' for each master. Also the core-dns docker image used was tagged with '1.6.7-1', apparently causing the corefile migration version check to fail

Anything else we need to know?:
This seems to be caused by

kubernetes/cmd/kubeadm/app/phases/addons/dns/dns.go

Line 453 in 39ed64e

patch := fmt.Sprintf(`{"spec":{"template":{"spec":{"volumes":[{"name": "config-volume", "configMap":{"name": "coredns", "items":[{"key": "%s", "path": "%s"}]}}]}}}}`, coreDNSCorefileName, coreDNSCorefileName)

Both 'key' and 'path' are set to 'Corefile-backup' when this function is called. Should the 'path' always stay as 'Corefile' because that is where the container expects to find the config file?

Environment:

• Kubernetes version (use kubectl version):
  Client Version: version.Info{Major:"1", Minor:"17", GitVersion:"v1.17.3", GitCommit:"06ad960bfd03b39c8310aaf92d1e7c12ce618213", GitTreeState:"clean", BuildDate:"2020-02-16T17:51:57Z", GoVersion:"go1.13.6", Compiler:"gc", Platform:"linux/amd64"}
  Server Version: version.Info{Major:"1", Minor:"17", GitVersion:"v1.17.3", GitCommit:"06ad960bfd03b39c8310aaf92d1e7c12ce618213", GitTreeState:"clean", BuildDate:"2020-02-16T17:47:13Z", GoVersion:"go1.13.6", Compiler:"gc", Platform:"linux/amd64"}

• Cloud provider or hardware configuration:
  Custom deployment in Openstack environment

• OS (e.g: cat /etc/os-release):
  NAME="SLES"
  VERSION="15-SP1"
  VERSION_ID="15.1"
  PRETTY_NAME="SUSE Linux Enterprise Server 15 SP1"

• Kernel (e.g. uname -a):
  4.12.14-197.29-default

• Install tools:
  kubeadm

• Network plugin and version (if this is a network-related bug):

• Others:

[anttitapio]

/sig cluster-lifecycle

[rajansandeep]

/assign

[neolit123]

/area kubeadm