This is the multi-page printable view of this section. Click here to print.
To upgrade Huawei CSI from 2.x to 4.5.0, uninstall it by referring to the user guide of the earlier version and install Huawei CSI by referring to Installing Huawei CSI Using Helm.
To upgrade Huawei CSI from 2.x or 3.x to 4.5.0, see Upgrading from 2.x or 3.x to 4.x.
To upgrade Huawei CSI from 4.x to 4.5.0, see Upgrading Huawei CSI on Kubernetes, OpenShift, and Tanzu.
Use a remote access tool, such as PuTTY, to log in to any master node in the Kubernetes cluster through the management IP address.
Go to the helm/esdk working directory. For the directory path, see Table 1.
cd helm/esdk
Run the following command to query the historical versions of the CSI services deployed using Helm.
helm history helm-huawei-csi -n huawei-csi
The following is an example of the command output.
REVISION UPDATED STATUS CHART APP VERSION DESCRIPTION
1 Mon Jan 8 04:15:40 2024 superseded esdk-4.4.0 4.4.0 Install complete
2 Mon Jan 8 04:16:12 2024 deployed esdk-4.5.0 4.5.0 Upgrade complete
Run the following command to roll back the CSI services to the specified version.
In the preceding command, revision-number indicates a version number queried in 3. For example, the version is 1.
helm rollback helm-huawei-csi -n huawei-csi 1
The following is an example of the command output. If Rollback was a success is displayed in the command output, the CSI services are successfully rolled back to the specified version.
Rollback was a success! Happy Helming!
This section describes how to upgrade Huawei CSI.
During the upgrade or rollback, the existing resources such as PVCs, snapshots, and Pods will run properly and will not affect your service access.
- Some CSI 2.x versions are unavailable now. If the upgrade fails, CSI may fail to be rolled back to a version which is unavailable now.
- After an upgrade from 2.x, 3.x, or 4.x to 4.5.0, a Pod that has been provisioned in the source version may fail to be mounted again. For details, see Upgrading from 2.x or 3.x to 4.x.
- During the upgrade or rollback, you cannot use Huawei CSI to create new resources or mount or unmount an existing PVC.
- During the upgrade or rollback, do not uninstall the snapshot-dependent component service.
In CSI 2.x or 3.x, when block storage is used, the mapping with storage is set up in the huawei-csi-node service. Therefore, the huawei-csi-node service needs to communicate with the storage management network. Because the huawei-csi-node service is deployed as a DaemonSet, the huawei-csi-node service is deployed on each node in the cluster. As a result, in a large-scale cluster, each huawei-csi-node service sends requests to the storage and the number of storage connections may be fully occupied. Accordingly, huawei-csi-node cannot provide services properly. In CSI 4.x, the deployment model is optimized. The setup of the mapping with storage is migrated to the huawei-csi-controller service and the huawei-csi-node service does not need to communicate with the storage management network. This reduces the networking complexity of Huawei CSI. In addition, the huawei-csi-controller service is deployed as a Deployment. The number of copies is set based on the customer’s reliability requirements. Generally, the number of copies ranges from 1 to 3. Therefore, the number of connections between Huawei CSI and storage is greatly reduced, so that Huawei CSI can connect to a large-scale cluster. This change may cause a problem. That is, if a new mount process is generated after CSI is upgraded to 4.x but with workloads provisioned using 2.x or 3.x and the Container Orchestration (CO) system does not invoke the huawei-csi-controller service provided by Huawei CSI, the mounting will fail. For details, see A Pod Fails to Be Created and Message “publishInfo doesn’t exist” Is Displayed in the Events Log.
If you have evaluated the risks mentioned in the preceding notice and need to upgrade CSI from 2.x or 3.x to 4.5.0, perform the following steps to back up storage backend configurations:
Use a remote access tool, such as PuTTY, to log in to any master node in the Kubernetes cluster through the management IP address.
Run the following command to back up the backend information to the configmap.json file. For the OpenShift platform, replace kubectl with oc.
kubectl get cm huawei-csi-configmap -n huawei-csi -o json > configmap.json
Perform the upgrade according to the procedure described in Upgrading Huawei CSI.
Configure the storage backend by following the instructions in Managing Storage Backends according to the backend information backed up in Backing Up Storage Backend Configurations. After the storage backend is successfully configured, perform operations according to the risk handling methods described in the preceding notice to prevent problems during Pod failover.
If CSI of an earlier version is deployed using Helm, perform the following steps to upgrade Huawei CSI.
Use a remote access tool, such as PuTTY, to log in to any master node in the Kubernetes cluster through the management IP address.
Copy the CSI component package of the target version to any directory on the master node.
Go to the helm/esdk working directory. For the directory path, see Table 1.
cd helm/esdk
Run the kubectl apply -f ./crds/backend/ command to update the storage backend CRD.
kubectl apply -f ./crds/backend/
(Optional) Check snapshot-dependent components by following the instructions provided in Checking Volume Snapshot-Dependent Components. After confirming that the components are correct, run the kubectl apply -f ./crds/snapshot-crds/ –validate=false command to update the snapshot CRD. If controller.snapshot.enabled is set to false or the Kubernetes version is earlier than v1.17, you can skip this step. For details, see Table 2.
kubectl apply -f ./crds/snapshot-crds/ --validate=false
Run the following command to obtain the original service configuration file. helm-huawei-csi indicates the Helm chart name specified during the installation of the earlier version, and huawei-csi indicates the Helm chart namespace specified during the installation of the earlier version.
helm get values helm-huawei-csi -n huawei-csi -a > ./update-values.yaml
Run the vi update-values.yaml command to open the file obtained in 6, modify the images configuration items, and update the image to the latest version. For details about the parameters to be modified, see Table 1.
Table 1 images configuration items
Image used by Huawei backends to manage storageBackendContent resources. | ||
Image used by Huawei backends to manage storageBackendClaim resources. | ||
livenessprobe sidecar image. | ||
csi-provisioner sidecar image. | ||
csi-attacher sidecar image. | ||
csi-resizer sidecar image. | ||
csi-snapshotter sidecar image. | ||
snapshot-controller sidecar image. | ||
csi-node-driver-registrar sidecar image. |
(Optional) If you need to update configuration items or add configuration information during the upgrade, modify the configuration information in the update-values.yaml file by referring to Parameters in the values.yaml File of Helm.
During the upgrade, if the update-values.yaml and values.yaml configuration files contain the same configuration item, the configuration in the update-values.yaml file takes effect preferentially.
Run the following command to upgrade Huawei CSI. In the following command, helm-huawei-csi indicates the specified Helm chart name, huawei-csi indicates the specified Helm chart namespace, and update-values.yaml indicates the file obtained in 6.
helm upgrade helm-huawei-csi ./ -n huawei-csi -f ./values.yaml -f ./update-values.yaml
After the huawei-csi service is deployed, run the following command to check whether the service is started.
kubectl get pod -n huawei-csi
The following is an example of the command output. If the Pod status is Running, the service is started successfully.
NAME READY STATUS RESTARTS AGE
huawei-csi-controller-6dfcc4b79f-9vjtq 9/9 Running 0 24m
huawei-csi-controller-6dfcc4b79f-csphc 9/9 Running 0 24m
huawei-csi-node-g6f4k 3/3 Running 0 20m
huawei-csi-node-tqs87 3/3 Running 0 20m
You have downloaded the CSI software package of a new version.
- During the upgrade or rollback, the existing resources such as PVCs, snapshots, and Pods will run properly and will not affect your service access.
- During the upgrade or rollback, you cannot use Huawei CSI to create new resources or mount or unmount an existing PVC.
- During the upgrade or rollback, do not uninstall the snapshot-dependent component service.
You have downloaded the CSI software package of the source version.