woeful patrol
Hi,
we have installed the Trident Protect Helm chart (version 100.2510.0) with the correspoding clusterName parameter:
helm install trident-protect netapp-trident-protect/trident-protect --set clusterName=lab-acp --version 100.2510.0 --namespace trident-protect
Backup / Restore procedures work fine across namespaces and clusters, however we are missing the cluster name in tridentctl-protect when quering for snapshots / backups (see attachment).
To our understanding the trident protect controller manager pod, should auto populate the cluster name, when defined during the helm chart deployment.
Debugging the helm chart, we can see that this value is used in the trident-protect-env-config configmap and checking it in our cluster reveals it is correctly set:
> oc get cm trident-protect-env-config -n trident-protect -o yaml
apiVersion: v1
kind: ConfigMap
metadata:
name: trident-protect-env-config
namespace: trident-protect
data:
...
TRIDENT_PROTECT_CLUSTER_NAME: lab-acp
We then debugged the appvault S3 bucket to detect, which data was written on backup / snapshot procedures and discovered, that each snapshot / backup directory of an application has a corresponding snapshot / backup.json that contains a json formatted string of the resource that triggered the action. These also included information about an empty cluster-name label.
Therefore we tried to set a cluster-name label under metadata.labels to each backup / snapshot resource, resulting in a correctly written cluster name column in tridentctl-protect.
Questions
- Is our expectation correct, that the the helm chart parameter clusterName should automatically define a cluster name for each snapshot / backup action without specifying it on each snapshot / backup resource?
- While this workaround (defining cluster-name under metadata.labels) works for manual snapshot / backup actions, it does not work for the schedule resource. How can we work around this?
Thanks and BR,
Philipp
noble pollen