Closed ldominguezBold closed 1 year ago
your command shows: --cluster-reset-restore-path=/var/lib/rancher/rke2/server/db/snapshots/etcd-snapshot-******
but the error says: "snapshot-file-path":"/var/lib/rancher/rke2/server/db/etcd-tmp/member/snap/00000000005a0964.snap.db
Can you confirm that you are providing the path to an actual scheduled or on-demand etcd snapshot, and not one of the temporary WAL checkpoint snapshot files?
Yeap i can confirm that is one of the scheduled snapshot, not a checkpoint wal
even that file (without the .db) exists in that directory
Can you try updating to the most recent v1.24 release, and try the same process? If it continues to fail, you may want to try running rke2-killall.sh
and removing the old database files (/var/lib/rancher/k3s/server/db/etcd
) prior to running the restore. It seems like perhaps the current etcd datastore is corrupt, and that is preventing it from restoring?
Can you also confirm that rke2 IS NOT currently running when you are doing this?
I will try removing the etcd data dir, because the other actions I was tried already.
The solution was remove the old etcd
data directory, thanks.
Environmental Info: RKE2 Version:
Node(s) CPU architecture, OS, and Version:
Cluster Configuration:
1 server
Describe the bug:
I tried to restore a snapshot with this command:
Steps To Reproduce:
Expected behavior:
Expected that restore end like docs say.
Actual behavior:
The process end with an error:
Additional context / logs: