A new variable glusterfs_clear_peers is being added, which controlls
if peer reset will be executed.
Under peer reset we mean finding up a peer UUID which is part of the
cluster, but this UUID is no longer present in
`glusterfs_server_group_name` group.
If such peer is detected, we remove bricks and peer itself from the
cluster.
It should be safe to run with `glusterfs_clear_peers: true` but it is
set to `false` by default to reduce role runtime.
Change-Id: I6d44028fe8ecf9b2866bb04730003b3bb234cc66
Signed-off-by: Dmitriy Rabotyagov <[email protected]>
Dmitriy Rabotyagov
2c568e3f68