Lock required in delete share/snapshot

[Peter Wang]

Recently I was running tempest for Manila,
I noticed some times, Test cases issued 2 delete snapshot command nearly in the same time and
these 2 api call came to our backend driver in a row within 0.6 s, below the manila api/driver call sequence in time line:

18:20:56.264 DELETE_1 cames via HTTP DELETE
18:20:56.86 DELETE_2 cames vai HTTP POST(Action body: {"os-force_delete": null})
18:20:57.414 DELETE_1 finisned in driver
18:20:57.474 DELETE_2 finisned in driver

that's say, driver will received 2nd delete_snapshot request while processing the 1st delete_snapshot.
while from Cinder, cinder manager will not dispatch the 2nd delete_snapshot request to driver until 1nd delete_snapshot finished ,since cinder add @locked_snapshot_operation for delete_snapshot

my question is : can Manila also add lock as cinder do?

Thanks
Peter

[OpenStack Infra (hudson-openstack)]

Related fix proposed to branch: master
Review: https://review.openstack.org/289441

[Jason Grosso (jgrosso)]

is this still an issue that is being worked on? Is it still happening in the latest releases of Manila.

[Jason Grosso (jgrosso)]

If this is still occurring please open a new issue.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Goutham Pacha Ravi (<email address hidden>) on branch: master
Review: https://review.opendev.org/289441
Reason: No update for several releases. Please restore if you intend to drive this