Skip to content

Commit

Permalink
Updating Authorization documentation (#972) (#973)
Browse files Browse the repository at this point in the history
* updating generating token

* updating generating token

* updating generating token doc

* updating troubleshooting page
  • Loading branch information
PeresKereotubo authored Feb 7, 2024
1 parent 130d2ae commit 75628be
Show file tree
Hide file tree
Showing 2 changed files with 34 additions and 4 deletions.
5 changes: 5 additions & 0 deletions content/docs/authorization/cli.md
Original file line number Diff line number Diff line change
Expand Up @@ -247,6 +247,11 @@ Usually, you will want to pipe the output to kubectl to apply the secret
```bash
karavictl generate token --tenant Alice --admin-token admintoken.yaml --addr csm-authorization.host.com | kubectl apply -f -
```
The token is read once when the driver pods are started and is not dynamically updated. If you are applying a new token in an existing driver installation, restart the driver pods for the new token to take effect.
```bash
kubectl -n <driver-namespace> rollout restart deploy/<driver>-controller
kubectl -n <driver-namespace> rollout restart ds/<driver>-node
```

### karavictl role

Expand Down
33 changes: 29 additions & 4 deletions content/docs/authorization/troubleshooting.md
Original file line number Diff line number Diff line change
Expand Up @@ -15,10 +15,7 @@ The CSM Authorization RPM will be deprecated in a future release. It is highly r
- [Running `karavictl tenant` commands result in an HTTP 504 error](#running-karavictl-tenant-commands-result-in-an-http-504-error)
- [Installation fails to install policies](#installation-fails-to-install-policies)
- [After installation, the create-pvc Pod is in an Error state](#after-installation-the-create-pvc-pod-is-in-an-error-state)

## Helm Deployment
- [The CSI Driver for Dell PowerFlex v2.3.0 is in an Error or CrashLoopBackoff state due to "request denied for path" errors](#the-csi-driver-for-dell-powerflex-v230-is-in-an-error-or-crashloopbackoff-state-due-to-request-denied-for-path-errors)

- [Intermittent 401 issues with generated token](#intermittent-401-issues-with-generated-token)
---

### The Failure of Building an Authorization RPM
Expand Down Expand Up @@ -97,6 +94,23 @@ Run the following commands to allow the PVC to be created:
semanage fcontext -a -t container_file_t "/var/lib/rancher/k3s/storage(/.*)?"
restorecon -R /var/lib/rancher/k3s/storage/
```
### Intermittent 401 issues with generated token
This issue occurs when a new access token is generated in an existing driver installation.

__Resolution__

If you are applying a new token in an existing driver installation, restart the driver pods for the new token to take effect. The token is read once when the driver pods are started and is not dynamically updated.
```bash
kubectl -n <driver-namespace> rollout restart deploy/<driver>-controller
kubectl -n <driver-namespace> rollout restart ds/<driver>-node
```

## Helm Deployment
- [The CSI Driver for Dell PowerFlex v2.3.0 is in an Error or CrashLoopBackoff state due to "request denied for path" errors](#the-csi-driver-for-dell-powerflex-v230-is-in-an-error-or-crashloopbackoff-state-due-to-request-denied-for-path-errors)
- [Intermittent 401 issues with generated token](#intermittent-401-issues-with-generated-token)

---


### The CSI Driver for Dell PowerFlex v2.3.0 is in an Error or CrashLoopBackoff state due to "request denied for path" errors
The vxflexos-controller pods will have logs similar to:
Expand Down Expand Up @@ -184,3 +198,14 @@ kubectl -n <namespace> rollout restart deploy/proxy-server
kubectl -n <driver-namespace> rollout restart deploy/vxflexos-controller
kubectl -n <driver-namespace> rollout restart daemonSet/vxflexos-node
```

### Intermittent 401 issues with generated token
This issue occurs when a new access token is generated in an existing driver installation.

__Resolution__

If you are applying a new token in an existing driver installation, restart the driver pods for the new token to take effect. The token is read once when the driver pods are started and is not dynamically updated.
```bash
kubectl -n <driver-namespace> rollout restart deploy/<driver>-controller
kubectl -n <driver-namespace> rollout restart ds/<driver>-node
```

0 comments on commit 75628be

Please sign in to comment.