d153ff2621
3PAR driver creates a dependency chain when we create a volume from a snapshot and while deleting the snapshot, it renames the dependent volume and also changes some of it's properties to break the dependency to finally delete the snapshot successfully. During the renaming, 3PAR driver creates a new volume named "omv-<id>" which is different from the original volume name "osv-<id>". NOTE: here "id" doesn't refer to OpenStack resource/volume ID but rather the ID provided by backend. We can fail to create the new volume for certain reasons and one of the reason is there already exists an internal volume named "omv-<id>" Since the IDs are unique, this volume could be a side effect of a failed operation earlier and the volume was not cleaned up. The ERROR log in this failure case shows the original volume name "osv-<id>" with a Duplicate exception[1] which doesn't provide any insight of the failure since we already know "osv-<id>" exists in backend and we are failing due to "omv-<id>" volume which needs to be manually cleaned up from 3PAR array. This patch corrects the volume name in the above failure case. [1] ERROR oslo_messaging.rpc.server cinder.exception.Duplicate: Volume (osv-<id>) already exists on array. Change-Id: I118339e87eefca2f0df90992c81d4f09fb6d0b68 |
||
---|---|---|
.. | ||
__init__.py | ||
hpe_3par_base.py | ||
hpe_3par_common.py | ||
hpe_3par_fc.py | ||
hpe_3par_iscsi.py | ||
nimble.py |