Hi, one doubt regarding etcd restoration , in lecture it is mentioned to stop ku . . .

Raj:
Hi, one doubt regarding etcd restoration , in lecture it is mentioned to stop kube api service first. In case cluster is deployed using kubeadm then I guess kube api will exist as a pod instead of service. Can you confirm how to deal in that situation ?

Memo Toro:
Hi @Raj for the purpose of the exam, you don’t need to stop the api service. The comment about stop api is more for production environments were you don’t want your backup to be out of sync since the moment you execute the command create the backup and until it creates the backup file. During that time your api could receive more data from other places to changes the state. Exam-wise not a problem. Let me know if it’s clear.

Raj:
Perfectly make sense , thanks for prompt reply

Raj:
I just tried to follow the steps in lab env but seems it got messed up, can someone suggest what’s wrong here :

Raj:

Raj:

Marc Richardson:
your instance looks horked, can you reboot it? Failed to connect to bus is an OS level problem, not kubernetes

Raj:
Thanks for your response , it was on test labs, per events it seems something with master node itself. Moreover lab expires in 60 minutes so don’t had much control on it.