Currently, as part of system restore, maintenance will lock all the
nodes except for controller-0. When locking storage nodes if the
replication will be lost, we query ceph pools to see if they are
empty. But during restore, ceph cluster is not up yet; therefore
query ceph pools will fail and it will block locking storage nodes.
Solution: In the case of the restore, if it is a force-lock, we
should skip the ceph pool query and just lock the storage node.
Reviewed: https:/ /review. openstack. org/612422 /git.openstack. org/cgit/ openstack/ stx-config/ commit/ ?id=5e2772b25e9 2730e9be990f57e 9aab57c3412135
Committed: https:/
Submitter: Zuul
Branch: master
commit 5e2772b25e92730 e9be990f57e9aab 57c3412135
Author: Wei Zhou <email address hidden>
Date: Mon Oct 22 10:55:32 2018 -0400
Lock storage nodes during system restore
Currently, as part of system restore, maintenance will lock all the
nodes except for controller-0. When locking storage nodes if the
replication will be lost, we query ceph pools to see if they are
empty. But during restore, ceph cluster is not up yet; therefore
query ceph pools will fail and it will block locking storage nodes.
Solution: In the case of the restore, if it is a force-lock, we
should skip the ceph pool query and just lock the storage node.
Change-Id: I0501b44ebe635f 36d7401437bfe28 9de7ee9fc73
Closes-Bug: 1798197
Signed-off-by: Wei Zhou <email address hidden>