[Enhancement] Analyse the active snapshot upload when corresponding etcd shutdown/crashes.

[ishan16696]

Feature (What you would like to be added):
Analyse the behaviour of etcd-backup-restore during active snapshot upload and when its corresponding etcd process shutdown/crashes.

Motivation (Why is this needed?):
Currently when corresponding etcd shutdown/crashes while backup-restore is uploading the snapshot, snapshotter wait for context timeout set by flag --etcd-snapshot-timeout (15mins in our production) untill current snapshot upload is finished.

• In case of single node etcd,
  • If etcd restarts, then backup-restore will starts the initialisation phase while full-snapshot is currently being uploaded.
• In case of multi-node etcd,
  • If etcd leader restarts, then it might possible that it will become follower(lost leadership) and other cluster member will become cluster leader, hence there will be new backup-leader who will also try to take snapshot while full-snapshot is currently being uploaded by previous backup-leader.