cloud-provider-openstack [cinder-csi-plugin] node server should not use openstack credentials by default

Is this a BUG REPORT or FEATURE REQUEST?:

/kind feature

What happened:

Cinder CSI node server tries to create an openstack client based on cloud-config, when usually this is not necessary, since all openstack related communications are done on csi controller side.

What you expected to happen:

I expect CSI cinder node server not to use openstack credentials, when it is not necessary and respect the https://github.com/kubernetes/kubernetes/blob/08e1fd3bb947faf465e8a67d5c7106dbd10840c0/pkg/apis/core/types.go#L1607 and https://github.com/kubernetes/kubernetes/blob/08e1fd3bb947faf465e8a67d5c7106dbd10840c0/pkg/apis/core/types.go#L1664..L1670 API. So when ephemeral storage is requested, user must provide secret reference via pod manifest or storage manifest.

In addition, node server manifest should not refer to secrets. The node server configuration should be done via configMap. See discussion here: https://github.com/kubernetes/cloud-provider-openstack/pull/861#issuecomment-585645374

Anything else we need to know?:

https://github.com/kubernetes-csi/external-attacher/pull/38

https://kubernetes-csi.github.io/docs/secrets-and-credentials.html

openstack client is used in functions listed below:

nodePublishEphermeral
nodeUnpublishEphermeral
NodeStageVolume
NodeUnstageVolume
NodePublishVolume
NodeUnpublishVolume

/cc @rfranzke

Apr 09 '20 15:04 kayrus

/cc @adisky

Apr 09 '20 15:04 rfranzke

just to confirm, the overall idea is to use secret or related info to avoid usage of cloud-config, but still need the provided secret to do communication with openstck side(in other words ,the mechanism of communication between CSI node + openstack is not changed, how to provide cloud definition will be changed, correct?)

Apr 13 '20 02:04 jichenjc

Initial idea was to start nodeserver without secrets. It may be required by security purpose. Additionally, follow CSI spec and configure openstack client from secrets, provided in storageclass manifest when possible.

For now I'm confident that secrets are required by ephemeral inline storage. Nodeserver may work fine without secrets and ephemeral storage support.

The rest functions are:

NodeStageVolume
NodeUnstageVolume
NodePublishVolume
NodeUnpublishVolume

I haven't managed to clarify their purpose. But it would be fine if it is nodeserver just return an error, when openstack API operation is not possible.

Apr 13 '20 06:04 kayrus

@kayrus we need to use openstack credentials for some of the operations in nodeserver . As per the https://kubernetes-csi.github.io/docs/secrets-and-credentials.html , secret is required at the "per driver" granularity (not different "per CSI operation" or "per volume"), then the secret SHOULD be injected directly in to CSI driver pods via standard Kubernetes secret distribution mechanisms during deployment. This is the way we are doing it currently , so it is fine.

IIUC, secrets needs to be provided per storageclass only when needed credentials per volume/per operation If a CSI Driver requires secrets "per CSI operation" or "per volume" or "per storage pool", the CSI spec allows secrets to be passed in for various CSI operations. Cluster admins can populate such secrets by creating Kubernetes Secret objects and specifying the keys in the StorageClass or SnapshotClass objects (Ref: https://kubernetes-csi.github.io/docs/secrets-and-credentials.html)

updating slack discussion on questions you asked:

do you agree that nodeserver must start without secrets? ramineni: we could explore this option, more investigation needed if this can be done without breaking the driver.
do you agree that cloud-config should be split into two parts: config and secret? ramineni: yes, this would make sense

Apr 13 '20 09:04 ramineni

Issues go stale after 90d of inactivity. Mark the issue as fresh with /remove-lifecycle stale. Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta. /lifecycle stale

Jul 12 '20 09:07 fejta-bot

/remove-lifecycle stale

Jul 13 '20 04:07 ramineni

Issues go stale after 90d of inactivity. Mark the issue as fresh with /remove-lifecycle stale. Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta. /lifecycle stale

Oct 11 '20 04:10 fejta-bot

/remove-lifecycle stale

Oct 12 '20 04:10 ramineni

Issues go stale after 90d of inactivity. Mark the issue as fresh with /remove-lifecycle stale. Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta. /lifecycle stale

Jan 10 '21 04:01 fejta-bot

/remove-lifecycle stale

Jan 11 '21 05:01 ramineni

See an aws node controller for reference: https://github.com/kubernetes-sigs/aws-ebs-csi-driver/blob/master/deploy/kubernetes/base/node.yaml It doesn't have secrets

Feb 12 '21 13:02 kayrus

/assign Explore nodeserver to be started without need of secrets

May 11 '21 10:05 ramineni

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Aug 09 '21 10:08 k8s-triage-robot

/remove-lifecycle stale

Aug 23 '21 10:08 ramineni

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Nov 21 '21 10:11 k8s-triage-robot

/remove-lifecycle stale

Nov 24 '21 03:11 ramineni

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Feb 22 '22 04:02 k8s-triage-robot

/remove-lifecycle stale

Feb 22 '22 05:02 ramineni

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

May 23 '22 06:05 k8s-triage-robot

/remove-lifecycle stale

May 23 '22 06:05 jichenjc

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Aug 21 '22 07:08 k8s-triage-robot

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten

Sep 20 '22 07:09 k8s-triage-robot

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Reopen this issue with /reopen
Mark this issue as fresh with /remove-lifecycle rotten
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/close not-planned

Oct 20 '22 07:10 k8s-triage-robot

@k8s-triage-robot: Closing this issue, marking it as "Not Planned".

In response to this:

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.

This bot triages issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied

After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied

After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Reopen this issue with /reopen

Mark this issue as fresh with /remove-lifecycle rotten

Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/close not-planned

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.

Oct 20 '22 07:10 k8s-ci-robot

It looks to me like node servers needed secrets for ephemeral storage, but ephemeral storage was removed in https://github.com/openshift/openstack-cinder-csi-driver-operator/pull/76 . Does this mean the requirement to have secrets on the nodes is gone?

Jul 20 '23 04:07 puck