Upgrade the cluster
Before upgrading the kubernetes cluster, ensure that the NetBackup Catalog is successfully backed up on cloud LSU (MSDP-C).
To upgrade AKS or EKS cluster hosting Cloud Scale
- Pause NetBackup job scheduling and allow all running jobs to be completed.
- Stop the primary services as follows:
For NetBackup version 10.5 and above, run the following script with a stop option:
/VRTSk8s-netbackup-<version>/scripts$ ./cloudscale_restart.sh stop
For NetBackup version lower than 10.5:
Get the environment name using the following command:
kubectl get environment -n <namespace>
Pause using the following command:
kubectl patch environment <env_name> -n <namespace>--type=merge -p='{"spec": {"primary": {"paused": true}}}
Scale down the primary pod to 0 as follows:
Get primary sts name: kubectl get sts -n <namespace> | grep "\-primary"
Scale to 0: kubectl scale --replicas=0 sts <sts name> -n <namespace>
- Upgrade cluster version from AKS/EKS portal.
- For EKS, update the add-ons by going to the add-on section (new defaults which are shown).
If any problem occurs during the add-on upgrade, use the following command if there is no customization made as part of the add-on upgrade:
aws eks update-addon \ --cluster-name <cluster_name> \ --addon-name <addon_name> \ --addon-version <addon_version> \ --resolve-conflicts OVERWRITE
- Upgrade all node pools version. This can be done in parallel with a console.
- For multiple cluster version upgrades, repeat steps 3 to 5.
- Once the cluster, the add-on and the node pool are upgraded to the required version, perform the following:
For NetBackup version 10.5 and above: run the following command to restart:
./cloudscale_restart.sh start
For NetBackup version lower than 10.5:
kubectl patch environment <env_name> -n <namespace>--type=merge -p='{"spec": {"primary": {"paused": true}}}
- Resume NetBackup job scheduling.