PowerProtect Data Manager 19.16 Kubernetes User Guide

PDF

Controller Configurations

Within the PowerProtect Data Manager UI, you can add controller configurations for a Kubernetes cluster.

You can specify the following controller configuration keys when adding or editing the Kubernetes cluster asset source Advanced Options.

Table 1. Controller Configuration KeysController configuration keys
Key	Value	Description	Examples
`k8s.docker.registry`	`fqdn`:`port`	The fully qualified domain name (FQDN) of the internal registry for air gapped clusters.	`artifacts.example.com` or `artifacts.example.com:8446`
`k8s.image.pullsecrets`	`pull secret name`	Name of the pull secret image that you create in the powerprotect namespace and the velero-ppdm namespace when containers are hosted in an internal registry and the registry requires authentication.	`kubernetes.io/dockerconfigjson`
`k8s.ppdm.autoenable.cbt`	true or false	This setting is only applicable for PVCs provisioned by the VMware CSI driver. PowerProtect Data Manager enables changed block tracking (CBT) by default on First Class Disks (FCDs). Set the value to `false` to disable this behavior and perform only full backups.
`k8s.ppdm.csi.nonsnapshot.storageclasses`	`nonsnapshot_storageclass_1`, `nonsnapshot_storageclass_2`	A comma-separated list of nonsnapshot CSI storage class names. Use this key if the CSI driver does not support snapshots.	`vmware-file-sc`
`k8s.ppdm.create.timeout`	`timeout in seconds`	The length of time (in seconds) that the powerprotect-controller waits for the CSI snapshot and temporary PVC to be created during the backup before timing out. The default value is `600s`.	`600s`
`k8s.ppdm.readonly.storageclass`	`storageclass1`,`storageclass2`	A comma-separated list of storageclass names. During the backup for these storage classes, PowerProtect Data Manager uses ReadOnlyMany (ROX) mode to create a temporary PVC from the snapshot. NOTE:The provisioner for these storage classes should support ROX mode.	`standard`, `gold`
`k8s.ppdm.support.volumeGroup`	true or false	PowerProtect Data Manager uses volumegroup snapshots by default for PVCs provisioned by the PowerFlex CSI driver that have volume group snapshotting enabled. Set the value to `false` if you do not want to use volumegroupsnapshot.
`k8s.velero.exclude.resources`	`resourcetype1`,`resourcetype2`	A comma-separated list of Kubernetes resource types to be excluded for all backups in the Kubernetes cluster.	Run kubectl get all to retrieve a list of pods, services, daemon sets, deployments, replica sets, jobs, and statefulsets in a namespace, or run the following command to retrieve all resource types: `$ kubectl api-resources --verbs=list --namespaced -o name \| xargs -n 1 kubectl get --ignore-not-found --show-kind -n namespace`
`k8s.ppdm.vspherecsi.use.fsagent`	true or false	When set to `true`, allows data mover selection for PVCs provisioned by vSphere CSI. You can select the File System agent instead of the default VDDK by setting this parameter to `true`. When not set or set to `false`, VDDK is used. NOTE:Enabling this controller configuration key requires vSphere CSI Driver version 2.5.0 or later.
`ppdm.backup.concurrency`	`number of concurrent namespace backup jobs`	The number of concurrent namespace backup jobs per cluster. The default value is `5`, the minimum value is `1`, and the maximum value is `50`. It can take up to 15 minutes for a change in the value to take effect. NOTE:Increasing the value increases the load on the Kubernetes cluster.	`5`
`k8s.ppdm.enable.parallel.pvc.backup`	true or false	When set to `true`, up to 5 PVCs will be backed up in parallel per namespace. For example, if `ppdm.backup.concurrency` is set to the default value of `5`, up to 25 PVCs will be backed up in parallel. The concurrent backup limit for PVCs per cluster is 50. NOTE:Increasing this value along with setting `ppdm.backup.concurrency` increases the load on the Kubernetes cluster. When this feature is enabled for non-vSphere CSI volumes, it may increase the load on back-end storage.
`ppdm.restore.concurrency`	`number of concurrent namespace restore jobs`	The number of concurrent namespace restore jobs per cluster. The default value is `2`, the minimum value is `1`, and the maximum value is `20`. It can take up to 15 minutes for a change in the value to take effect. NOTE:Increasing the value increases the load on the Kubernetes cluster.	`2`

NOTE: When updates to Advanced Options result in changes to the configuration of PowerProtect Data Manager components in the Kubernetes cluster, the interruption of running Kubernetes protection activities can occur.