delete huge number of resources cluster might make resource service hang

[Rico Lin]

We got some report from Sahara and Magnum team when deleting a huge number of Cinder Volumes within resource group, might encounter with Cinder API leak and makes action stuck in in progress.
Actually to request for huge number action always become a burden for any services.
We should make it possible to create a delay when resource cluster deletion

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/501276

[Spyros Trigazis (strigazi)]

The problem in our deployment was in cinder. Even without heat we could make cinder fail by deleting 10 volumes at once. We have cinder since a while and the db connection wasn't configured to use pymyql which is the recommended one since a while.

Switching cinder to use pymysql solved the problem. Now we can delete without any problems 100s of volumes with or without heat.

[Zane Bitter (zaneb)]

The proposed patch requires a resource schema change and we are past feature-freeze, so bumping this to Rocky.