Dell PowerMax - Online migration can fail

[Jean Pierre Roquesalane]

During live migration, a timeout can happen due to some slowness when executing specific functions.

In order to reduce the slowness on the Dell PowerMax backend driver and to reduce the risk of long traffic disruptions when multiple detach and attach requests are invoked, the renaming of the device when detaching can be skipped as it will be deleted afterwards.

The current logic of the delete_volume function can cause slowness and break online migration.
https://github.com/openstack/cinder/blob/299553a4fe281cde9b14da34a470dcdb3ed17cc0/cinder/volume/drivers/dell_emc/powermax/rest.py#L1267

Please note that it only affects driver versions up through Yoga as the change was done starting Zed.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/cinder/+/894922
Committed: https://opendev.org/openstack/cinder/commit/d02c88346dbb586f82de91804e4d1bda34308608
Submitter: "Zuul (22348)"
Branch: stable/yoga

commit d02c88346dbb586f82de91804e4d1bda34308608
Author: Jean-Pierre Roquesalane <email address hidden>
Date: Wed Sep 13 15:56:15 2023 +0200

    Dell PowerMax : Disabling the renaming of a volume before deletion

    This change is to speed up the deletion of a volume which doesn't
    require to rename the volume prior to delete it.
    This removal will allow to save 15s in terms of operation time
    and avoid unnecessary pausing time during migration.
    This affects only version up to Yoga as this was included into a
    bigger patch merged in Zed.
    This is just a one line removal and doesn't affect other stuff.

    Closes-Bug: #2033398
    Change-Id: I80088ba4171516d5848d510bbe986fb92a734e38

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/cinder 20.3.2 release.