| Version: | 1.4.0 |
|---|---|
| Revision: | Thursday Jan 14, 2021 |
The following table lists the resolved issues for HPE CSI Driver for Kubernetes v1.4.0.
The following table lists the known issues for HPE CSI Driver for Kubernetes v1.4.0. Please note that, Known Issues from previous releases are still applicable with suggested workarounds, if they are not part of Resolved Issues above.
| ID | Component | Title | Description | Workaround |
|---|---|---|---|---|
| CON-1175 | csi.k8s | Not able to create Inline volumes/PODs on SUSE Caasp setup | Pod spec needs to have serviceAccount with associated pod security policies (PSP) to allow “csi” volumes. | 1. Add "csi" to the "Privileged" PSP volume list. kubectl edit psp suse.caasp.psp.privileged 2. Create Service account & cluster bindings for pod-inline. 3. Add the service account created in step 2 in pod spec. Please refer to https://documentation.suse.com/suse-caasp/4.2/single-html/caasp-admin/index.html#_pod_security_policies. |
| CON-1476 | csi.k8s | PVCs parameters which are allowed in mutations and not in overrides are updated automatically | PVC with annotation for parameters which are allowed in "mutation" and not in "Overrides" gets updated automatically after about 2 minutes of time. Volumes get removed from volume collection as well | Mutation fields need to be present for overrides in storage class to prevent this error. |
| CON-1569 | csi.k8s | Restart of CSI Driver pods are needed when password is updated | When the user password is updated on the HPE Nimble Storage backend, the cached tokens do not expire causing the CSI driver to continue to work . | HPE CSI driver pods need to be restarted after making any change to the password. |
| CON-1630 | csi.k8s | Failover of PODs after changing CHAP password fails | After CHAP password update, the pods fails to to come up on the failover node | Sometimes, we have seen that the pods get created on the failover node first before getting deleted from the original node. To overcome this, cordon the original node and then perform a drain of the pods. |
| CON-1686 | csi.k8s | NodeGetVolumeStats fails for NFS Volumes | NFS Volumes have volumeHandle which do not map to actual CSP backed volumes. NodeGetVolumeStats will fail for these PVC and fill up the CSI Driver and CSP logs. | This is fixed with quay.io/hpestorage/csi-driver:v1.4.0-CON-1686 and will be officially available in the next release. |
| BZ-134925 | csi.k8s | Disable NodeGetVolumeStats capability | Primera/3PAR CSP is filled with getvolumebyid calls due to nodegetvolumestats capability which impacts attach/detach during drain of node on a scaled environment . | This is fixed with quay.io/hpestorage/csi-driver:v1.4.1-BZ-134925 and will be part of next release of csi-driver. Please update only the hpe-csi-node daemonset with the patch. |
The following table lists the known issues for HPE Primera/3PAR CSP for HPE CSI Driver release 1.4.0
| ID | Component | Title | Description | Workaround |
|---|---|---|---|---|
| CIDG-1928 | Volume Groups | Volume group deletion | Volume group can be deleted even though some PVC's are still associated to it | Due to a limitation in volume group sidecar, we are not able to prevent this operation from happening. It's better to check for the kubectl get events to understand the volume group deletion failures. Note: Volume group CR gets deleted but the a volume group content and the storage backed volume group remains |
| CIDG-1304 | Volume Mutator | Unset description | Currently we can set description field for the PVC, but if already set, we can't unset the same | None |
| CIDG-1305 | Volume Mutator | Error validation in multiple edits | If multiple fields of PVC are edited, and if one of them is incorrect, partial edits are not done, and it results in a failure. kubectl get events again can be used for detecting such failure |
Retry the operation of edit with all proper values by looking at the kubectl events logs |
| CIDG-1302 | Volume Mutator | Incorrect PVC edit | If there is a validation failure due to incorrect PVC edit, then the operation does not stop until the user corrects the edit operation. This is a limitation for now. | |
| CIDG-1327 | Replication | System limitation | Handling system limitation of RCG | If more than 60 volumes are added into a RCG (Remote Copy Group) as part of parallel volume addition, due to the stop/start of RCG internally by CSP, the backend array can reach system limits and can cause PVC creation failure. Limit the number of PVCs to be added |
| BZ-134504 | csi.k8s | More online FC ports caused huge time for mount | By default, the CSP creates vluns on all the FC ports that it sees online. In some environments, with large number of fc ports it leads to increased mount time. Added a new option to read a list of FC ports from env variable "FC_PORTS_LIST", to control what ports are considered for vlun creation. | This is fixed with quay.io/hpestorage/hpe3parprimera-csp:v1.2.1 and will be part of next release of csi-driver. FC_PORTS_LIST env variable needs to be added in primera3par-csp deployment. |