How to troubleshoot if restart deployment but nothing happens with no event, not . . .

Kubernetes-Slack Discussions
#21

K8s_Member:
My control plane component is not running as pod, it’s service on master node instead.

#22

Alex Novik:
systemctl status kube-controller-manager

#23

Alex Novik:
on your master node

#24

K8s_Member:
it’s running.

#25

Alex Novik:
check logs

#26

Alex Novik:
cat /var/log/kube-controller-manager.log

#27

K8s_Member:
Thank you, let me check

#28

Alex Novik:
deployments and rs are controlled by kube-controller-manager and its replication and deployment controllers, to if nothing happens when you try to deploy I suspect this one is the issue.

#29

K8s_Member:
I totally agree with you.
When any kubectl request is comming, kube-apiserver is responsible for create the resource then put them into queue. Controller manager like Deployment-Controller or Replicasets-controller will continuous monitoring to consume them on queue to take action, right?

#30

Alex Novik:
Yeah

#31

Alex Novik:
my action plan would look like:
• check the versions (make sure they are the same for all master node components
• check the logs for master-controller
• delete deployment and any existing RS associated with this Deployment if any
• restart master-controller service
• check etcd logs for any error codes related to master-controller