Procedure to rollback when upgrade of Snapshot Manager fails
If upgrade of NetBackup Snapshot Manager fails and the user wants to rollback to previous version (any version up to 10.1), perform the steps described below.
Procedure to rollback Snapshot Manager
- Make a copy of the
environment.yamlfile using the following command:cp environment.yaml environment_without_nbsm.yaml
- Remove the cpServer section from
environment_without_nbsm.yamlfile and apply it. This deletes the resources of cpServer.If any resources are present from flexsnap, then manually delete the resources using the following commands:
kubectl delete deployment <deployment name> -n <environment_namespace>
For example,
kubectl delete deployment.apps/flexsnap-mongodb -n nbux
kubectl delete job <job name> -n <environment_namespace>
(Applicable only if mongodb pvc exists)
kubectl get pvc -n <environment_namespace> | grep mongodb-pvc
If there is any response from the above command, then delete the psql pvc and pv. commands to delete the pvc
kubectl delete pvc psql-pvc -n <environment_namespace>
kubectl delete pv $(kubectl get pv | grep psql-pvc | awk '{printf $1" " }')
- Re apply the
environment.yamlfile to start reinstalling the cpServer resources:kubectl apply -f environment.yaml -n <environment_namespace>
- To validate the rollback for NetBackup Snapshot Manager, use one of the following methods:
Check the CRO status using the following command:
kubectl get cpservers -n <environment_namespace>
Use the following command to check the logs of each step after describing the above CRO:
kubectl describe cpservers <cpserver_cro_name> -n <environment_namespace>
For example, kubectl describe cpservers cpserver-1-224245 -n nbux
Note:
Ensure that all the pods are in running state.
- Use the following command to verify the operator logs:
kubectl logs -f <flexnsap_operator_pod_name> -n <netbackup-operator-system>
- After installation completion, describe any flexsnap pod to verify if it has the older tag.