David Brunkow:
Anybody have any good instructions for etcd backup and restore when your etcd is a static pod?
David Brunkow:
Should I assume that since the instructions on http://kubernetes.io|kubernetes.io only specify for a service based etcd that static pod etcd won’t be on the exam?
seand:
take a backup
move staticpod manifest to your -/
delete the data dir declared in the manifest
restore from backup passing certs and snapshot file
move manifest back
that will bring your backup live with the least amount of effort.
edit: on phone but I can be more explicit if needed:)
David Brunkow:
I go into the etcd pod, use etcdctl there and take a backup. exit the pod, move the etcd.yaml and the pod stops running, I no longer have access to etcdctl.
David Brunkow:
probably need to install etc-client on the control plain node, run the backup and restore from there, modify the etcd.yaml to the new path on the controlplain, have it restart the etcd pod pointing to the new path.
seand:
why are you going into that pod? it’s on master/controlplane so etcdctl is used from there.
I’ll post steps tomorrow if you’re still needing help… not home by my computer
David Brunkow:
figured it out. I assumed we wouldn’t have access to etcdctl so used in etcd pod. Instead, I installed etcd-client, ran etcdctl from controlplane. thanks