Need "force" option for gluster snapshot delete

[kohlschuetter]

When trying to delete snapshots (or an entire volume, for which snapshots need to be deleted first), there can be a situation where a failed brick cannot delete a snapshot (because the entire underlying brick file system is gone, for example).

A typical error message in this case is

snapshot delete: failed: Pre Validation failed on <NODE>. Please check log file for details.

In this situation, a "force" option for "gluster snapshot delete" would be desirable.

The only reliable workaround I found was to stop glusterd for the entire cluster, manually delete the metadata directory /var/lib/glusterd/snaps/<SNAPSHOTID> for the corresponding affected snapshot on all servers, then restarting glusterd.

This of course is an undesired downtime for all other, otherwise unaffected volumes.