Fix requeue logic in the common controller

[xing-yang]

What type of PR is this?
/kind bug

What this PR does / why we need it:

Which issue(s) this PR fixes:
Fixes #

Special notes for your reviewer:

Does this PR introduce a user-facing change?:

Fixes the requeue logic in the snapshot controller.

[xing-yang]

/assign @yuxiangqian

[yuxiangqian]

WIP

[xing-yang]

@yuxiangqian Addressed your comments. PTAL. Thanks.

[xing-yang]

/assign @bswartz

[xing-yang]

@yuxiangqian Addressed your comments. PTAL.

[bswartz]

Tested this PR, and it works as well or better than master in my torture tests.

[xing-yang]

Thanks @bswartz !

[msau42]

@bswartz side topic but is it possible to generalize your stress test and add it to the kubernetes storage e2e test suite? I think we don't do enough stress testing in k8s so this would be a great addition.

[bswartz]

@bswartz side topic but is it possible to generalize your stress test and add it to the kubernetes storage e2e test suite? I think we don't do enough stress testing in k8s so this would be a great addition.

My stress test is just an open-ended loop that creates volume from snapshots and snapshots from volumes, reading and writing data each time. I kick it off and let it run for an arbitrary period of time, sometime 5 minutes, sometimes 5 hours.

The pseudocode looks like this:

create_empty_volume(0) // creates empty volume 0
for i = 0; true; i++ {
create_pod_and_read_write_data(i) // creates a pod to read volume i, validate the data, and write some new data, and flush
create_snapshot(i) // create snapshot i from volume i
delete_pod(i)
delete_volume(i)
create_volume_from_snap(i+1) // creates volume i+1 from snapshot i
delete_snapshot(i)
}

After a few hours this loop can run hundreds of times. I was able to shake out some bugs in my CSI drivers using this approach because certain failures only manifested after a dozen or so cycles of create/delete.

I'm not sure how this can become an e2e test because of its open-ended nature. There's no obvious place to stop.

[msau42]

I've added a similar stress test (without snapshots) and made the parallelization and # iterations configurable:
https://github.com/kubernetes/kubernetes/blob/master/test/e2e/storage/testsuites/stress.go

[yuxiangqian]

/cc @AndiLi99

[k8s-ci-robot]

@yuxiangqian: GitHub didn't allow me to request PR reviews from the following users: andili99.

Note that only kubernetes-csi members and repo collaborators can review this PR, and authors cannot review their own PRs.

In response to this:

/cc @AndiLi99

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[yuxiangqian]

@xing-yang lgtm, please squash

[xing-yang]

@yuxiangqian Squashed commits. PTAL.

[yuxiangqian]

/lgtm

[yuxiangqian]

/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: xing-yang, yuxiangqian

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [xing-yang,yuxiangqian]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment