From external jump host that has access to cluster and etcd endpoints it seems to be possible to get a backup and also perform restore as per some conversations I have seen online, for backup if you have certificates and key in jump host you can point to etcd endpoint and run the snapshot save (correct me if I am wrong), but for the restore I don’t understand how this works since I assume that restore needs to be done locally in master node where etcd static pod is running or in external etcd host. I am confused, Can you please help me? I am about to take the exam but I am so worried about this backup/restore method since it is not clear where do you need to leave the backup (in exam)(master node, jump host??) and also restoration issues.
First, don’t worry about the exam; you need to understand the principles of doing a restore, but you won’t be asked to do a full etcd store during the exam – it’s too likely to cause trouble and break an exam cluster. When I’ve seen lab questions, you are typically told where to restore the data directory, but you don’t update etcd’s or kube-apiserver’s config files to point to the new directory. You certainly won’t be asked to do the latter on the exam.
As far as restoring etcd, you normally do this on the system that’s running etcd, since the restore creates a directory with the data files.