OSPCIX-1049 pointed out a number of manila tempest test failures in the RHOSO 18 Unibeta job, blocking promotions

A bug was identified with NetApp ONTAP where snapshot clones could not be immediately deleted. ONTAP API behavior around cloning has changed where if a flexclone is deleted as soon as it is created, the deletion API fails briefly. Tempest tests can run into this because a number of them:

• Create a share (a.k.a NetApp Flexvol)
• Create a snapshot of share
• Create a share from snapshot (a.k.a NetApp FlexClone)
• Delete these in reverse order

The deletion of the second share fails on ONTAP with a message asking for clones splitting operations to cease before attempting deletion. 

This bug has been reported upstream for NetApp to address. In the meantime, the following tests are skipped:

• manila_tempest_tests.tests.scenario.test_share_basic_ops.TestShareBasicOpsNFS.test_write_data_to_share_created_from_snapshot

and "[share]/suppress_errors_in_cleanup" has been set to true in tempest.conf. 

Doing this allows us to unblock CI while we await a fix; however, we run into teh risk of introducing a regression in the interim.

Bug impact

• It's unlikely that users create clones of snapshots and immediately attempt deleting the clone. 
• This seems to only be concerning in tempest test environments

Known workaround

• Shares are stuck in "error_deleting". Their status can be reset and deletion can be attempted again. Since the deletion is blocked because of a clone job that was in the process of being canceled, it's a timing issue - so in all probability, deletion would succeed if attempted again.