enable topology support

[mucahitkurt]

What type of PR is this?
/kind feature

What this PR does / why we need it:
Enable topology support of the driver

Which issue(s) this PR fixes:
Fixes #54

Special notes for your reviewer:
I have not tested well yet, opened the PR to get feedbacks if I miss anything, I will try to test the change.

Does this PR introduce a user-facing change?:

Enable topology support of CSI hostpath driver.

cc @msau42

[k8s-ci-robot]

Welcome @mucahitkurt!

It looks like this is your first PR to kubernetes-csi/csi-driver-host-path 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes-csi/csi-driver-host-path has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

[k8s-ci-robot]

Hi @mucahitkurt. Thanks for your PR.

I'm waiting for a kubernetes-csi or kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

[msau42]

/ok-to-test

[msau42]

CreateVolume also needs to be updated to return topology in the response

[msau42]

topology also has to be feature-gated on in the csi-provisioner spec

[mucahitkurt]

fixed, thanks!

[msau42]

sorry, I just realized that we can't update the deployment spec until we tag a new image. So we may need to split this into 2 PRs:

1. add topology to the driver. this may need to be manually tested with private image. After this is merged, then we can cut a new hostpath driver image.
2. update the deployment spec

[mucahitkurt]

I reverted the deployment spec changes, I'll try to test the driver on my local env.

[mucahitkurt]

I'm just trying to understand, why should we separate deployment spec changes and driver topology support changes?

[msau42]

Actually, maybe it won't cause issues. The pull jobs should be replacing hostpath driver image with a custom one built from the PR. But the CI jobs run with the hostpath driver image specified in the deployment spec, so they will start running with topology enabled in the provisioner, but the tagged driver doesn't have it enabled yet. I think that may actually be fine though because the provisioner checks the feature gate + if the driver supports the capability. So actually, maybe there isn't a problem.

[mucahitkurt]

Then I'll add the feature gate to the provisioner?

[msau42]

/assign

[mucahitkurt]

/retest

[mucahitkurt]

I'm trying to understand why these tests failed, I want to run the some of failed test on my local kubernetes with the changed csi hostpath driver but haven't succeeded yet.

[msau42]

It looks like the tests are failing during hostpath driver bringup (node-driver-registrar is crashing?) You can try running deploy/kubernetes-1.15/deploy-hostpath.sh on your local cluster with a private hostpath driver image

[mucahitkurt]

It looks like the tests are failing during hostpath driver bringup (node-driver-registrar is crashing?) You can try running deploy/kubernetes-1.15/deploy-hostpath.sh on your local cluster with a private hostpath driver image

For the tests pull-kubernetes-csi-csi-driver-host-path-1-15-on-kubernetes-1-15, I think test failed after the hostpath driver bring up because I can see hostpath driver is up before the tests are started from the build log.

But It seems that volume provisioning, attachment are successful, but mount operation failed during the tests are running, and there are events about node-driver crash.

I'm trying a similar approach on my local env., but couldn't create the same test results. I'm running csi hostpath plugin from a local image that I build, I can run this test successfully.

When I try to run a e2e test, I can get success like this one.

Is there any way to get more detail(logs) about why a pod is crashing from these CSI tests?

[msau42]

You can get driver logs at: https://gcsweb.k8s.io/gcs/kubernetes-jenkins/pr-logs/pull/kubernetes-csi_csi-driver-host-path/88/pull-kubernetes-csi-csi-driver-host-path-1-15-on-kubernetes-1-15/1165389466169249793/artifacts/

[mucahitkurt]

/test pull-kubernetes-csi-csi-driver-host-path-1-13-on-kubernetes-1-13

[msau42]

This is the error from node-driver-registrar: https://storage.googleapis.com/kubernetes-jenkins/pr-logs/pull/kubernetes-csi_csi-driver-host-path/88/pull-kubernetes-csi-csi-driver-host-path-1-15-on-kubernetes-1-15/1169361761179013123/artifacts/csi-hostpathplugin-0/node-driver-registrar.log

I0904 21:35:31.590723       1 main.go:87] Received NotifyRegistrationStatus call: &RegistrationStatus{PluginRegistered:false,Error:RegisterPlugin error -- plugin registration failed with err: error updating Node object with CSI driver node info: error updating node: timed out waiting for the condition; caused by: failed to patch status "{\"metadata\":{\"annotations\":{\"csi.volume.kubernetes.io/nodeid\":\"{\\\"hostpath.csi.k8s.io\\\":\\\"csi-prow-control-plane\\\"}\"},\"labels\":{\"node.kubernetes.io/topology.hostpath.csi.k8s.io/node\":\"csi-prow-control-plane\"}}}" for node "csi-prow-control-plane": Node "csi-prow-control-plane" is invalid: metadata.labels: Invalid value: "node.kubernetes.io/topology.hostpath.csi.k8s.io/node": a qualified name must consist of alphanumeric characters, '-', '_' or '.', and must start and end with an alphanumeric character (e.g. 'MyName',  or 'my.name',  or '123-abc', regex used for validation is '([A-Za-z0-9][-A-Za-z0-9_.]*)?[A-Za-z0-9]') with an optional DNS subdomain prefix and '/' (e.g. 'example.com/MyName'),}
E0904 21:35:31.591022       1 main.go:89] Registration process failed with error: RegisterPlugin error -- plugin registration failed with err: error updating Node object with CSI driver node info: error updating node: timed out waiting for the condition; caused by: failed to patch status "{\"metadata\":{\"annotations\":{\"csi.volume.kubernetes.io/nodeid\":\"{\\\"hostpath.csi.k8s.io\\\":\\\"csi-prow-control-plane\\\"}\"},\"labels\":{\"node.kubernetes.io/topology.hostpath.csi.k8s.io/node\":\"csi-prow-control-plane\"}}}" for node "csi-prow-control-plane": Node "csi-prow-control-plane" is invalid: metadata.labels: Invalid value: "node.kubernetes.io/topology.hostpath.csi.k8s.io/node": a qualified name must consist of alphanumeric characters, '-', '_' or '.', and must start and end with an alphanumeric character (e.g. 'MyName',  or 'my.name',  or '123-abc', regex used for validation is '([A-Za-z0-9][-A-Za-z0-9_.]*)?[A-Za-z0-9]') with an optional DNS subdomain prefix and '/' (e.g. 'example.com/MyName'), restarting registration container.

[msau42]

I would remove the "node.kubernetes.io/" prefix. I think that's what the error is saying.

[mucahitkurt]

It was like that but I added that prefix because my previous error message said it's forbidden to use that kind of label, so when I check the error message I found this check, and allowed prefixes.

I think current error message is because of usage two / characters, so I expect that if I change the label name to something like node.kubernetes.io/topology.hostpath.csi.k8s.io-node it'll pass.

Do you know a way that I can over forbidden check and use node label like topology.hostpath.csi.k8s.io/node?

I guess there is a way because I can use topology.hostpath.csi.k8s.io/node label on my one-node local cluster without any forbidden error message.

I0824 22:30:52.401362       1 main.go:87] Received NotifyRegistrationStatus call: &RegistrationStatus{PluginRegistered:false,Error:RegisterPlugin error -- plugin registration failed with err: error updating Node object with CSI driver node info: error updating node: timed out waiting for the condition; caused by: failed to patch status "{\"metadata\":{\"annotations\":{\"csi.volume.kubernetes.io/nodeid\":\"{\\\"hostpath.csi.k8s.io\\\":\\\"csi-prow-control-plane\\\"}\"},\"labels\":{\"topology.hostpath.csi.k8s.io/node\":\"csi-prow-control-plane\"}}}" for node "csi-prow-control-plane": nodes "csi-prow-control-plane" is forbidden: is not allowed to modify labels: topology.hostpath.csi.k8s.io/node,}
E0824 22:30:52.401510       1 main.go:89] Registration process failed with error: RegisterPlugin error -- plugin registration failed with err: error updating Node object with CSI driver node info: error updating node: timed out waiting for the condition; caused by: failed to patch status "{\"metadata\":{\"annotations\":{\"csi.volume.kubernetes.io/nodeid\":\"{\\\"hostpath.csi.k8s.io\\\":\\\"csi-prow-control-plane\\\"}\"},\"labels\":{\"topology.hostpath.csi.k8s.io/node\":\"csi-prow-control-plane\"}}}" for node "csi-prow-control-plane": nodes "csi-prow-control-plane" is forbidden: is not allowed to modify labels: topology.hostpath.csi.k8s.io/node, restarting registration container.

[msau42]

I believe the restriction is on labels with k8s.io or kubernetes.io. Try leaving that part out of the label and see if that works

[mucahitkurt]

I changed to topology.hostpath.csi/node, it works, thanks!

[msau42]

Can you also update the example pods under the examples/ directory? They currently have a nodeSelector to be on the same node where the hostpath driver is deployed, but technically with topology enabled, we don't need that anymore.

[msau42]

/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: msau42, mucahitkurt

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 [msau42]

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

[mucahitkurt]

Can you also update the example pods under the examples/ directory? They currently have a nodeSelector to be on the same node where the hostpath driver is deployed, but technically with topology enabled, we don't need that anymore.

Only examples/csi-app.yaml has pod affinity settings, I deleted it.

Also for the below comment in the issue, should I open an issue?

...Then applications no longer need to be scheduled manually onto the right node and we can remove that workaround from the E2E testing in Kubernetes....

[msau42]

/lgtm
Thanks! For the in-tree hostpath e2e changes, opening an new issue sounds good. We can update in-tree tests once we cut a new release of the driver.