Cluster Migration Wizard

The Cluster Migration Wizard allows you to create and edit Kubernetes migration jobs. It consists of several steps, not all of which will appear in all cases. You can jump back and forth between steps by clicking on the step names on the left-hand side of the wizard.

Defining a Kubernetes migration job

  1. In the Cluster step of the wizard you must select the source cluster for migration from the list of active clusters. Note that depending on how you reached the wizard, the source cluster may already be selected for you. In that case, the wizard will open on the next step. After choosing a cluster, click Next.

  2. In the Selections step, you choose exactly what will be copied to the destination cluster using the following options:

    Namespaces

    Select either Full Cluster or Select Namespaces. Choosing Select Namespaces will prompt you to select the specific namespaces to include in the copy operation. Choosing Full Cluster (the default) will include all namespaces. Choosing Full Cluster and then Exclude namespaces will allow you to exclude specific namespaces.

    Include all cluster-scoped resources

    If enabled, the system will attempt to migrate all cluster-scoped resources. Otherwise, only cluster-scoped resources associated with selected resources (e.g. PVs associated with PVCs) will be migrated. Full cluster migrations always migrate all cluster-scoped resources, but when only selected namespaces are being migrated this is optional. This option defaults to off. This option is only shown if full cluster migration has not been selected (i.e. namespaces are selected or excluded).

    Advanced Options

    Opening the Advanced options section will allow you to set the following options:

    Select resource types (optional)

    Allows you to copy only specific Kubernetes resources types.

    Select labels (optional)

    Enter key-value pairs to specify the labels for the objects that you wish to copy in the cluster.

    Tip

    If you enter multiple label selectors separated by spaces in the Select Labels field, the relationship between them is assumed to be logical AND. So if you enter a:b c:d, it means that a must equal b AND c must equal d in order for the selector to match.

    CloudCasa will also allow you to enter a single key with multiple comma-separated values, and the relationship between these is assumed to be logical OR. So entering a:b,c means that a must equal b OR c. These can be combined, so entering a:b,c x:y means that a must equal b OR c AND x must equal y. That’s equivalent to (a == b || a == c) && x == y for you C lovers out there. Note that OR is not currently possible between different keys.

    Snapshot PVs before backup where possible

    Create snapshots of PVs that support them, and copy from the snapshots for consistency. This should generally be left enabled (the default) unless you experience problems with snapshot creation.

    Back up directly from all PVs not snapshotted

    All PVs not snapshotted will be copied directly. This should generally be left enabled (the default) unless you do not wish to copy volumes which do not support snapshots.

    Configure backup method by storage class

    Allows you to manually select the copy method for each storage class defined in your cluster. The available methods are:

    • Read data from snapshot

    • Read data from PVC

    • Read data from underlying host volume

    • Skip

    Destination storage for copy

    This allows you to select the temporary object storage used in the migration. Choose one of the following options:

    • Inherit from cluster preferences - Use the default storage destination for the cluster.

    • CloudCasa Storage - Use CloudCasa Storage. Select the cloud provider and region.

    • User-provided Storage - Select one of the user-defined object storage endpoints that you registered in the Storage page.

    Concurrent PVs

    Select the maximum number of PVs that will be copied at once. The allowed range is between 1 and 16. The default is 2 PVs.

    Total concurrent files

    Sets the number of files that will be copied concurrently across ALL PVs.

    Max transfer rate per PV

    Specifies the maximum transfer rate for the data transfer between each PV and the destination storage during the migration job, in MB per second. Leave this field blank to set no limit. The default value is blank for “Unlimited”.

    Data mover pod timeout

    Time that the agent will wait for the data mover pod to start. If using snapshots, the timer will start once snapshot PVCs are bound. This should only need to be changed in special cases, for example when copying Longhorn volumes using snapshots.

    Data mover memory limit

    Controls the maximum amount of memory that the data mover pod will be permitted to allocate during the migration process. This can be set between 512 MB and 8 GB. Higher values may be required if large settings for PV and/or file parallelism are used. By default it is set to 1024 MB.

    Exclude unattached PVCs

    By default, CloudCasa will include unattached PVCs in the copy operation. Enabling this option will cause unattached PVCs to be excluded from the copy. Be aware that there is currently no option to restore unattached PVCs to the migration target.

    Enable storage class mapping for PV snapshots mounted during backup

    This option allows you to use different storage classes when PV snapshots are mounted for copying the data off during migration operations. This may be useful to, for example, indicate to your storage system that less replicas or different parameters should be used for these transient volumes than for normal production volumes. By default, the storage class of the original source volume will be used.

    Note: You must specify a storage class that uses the same CSI driver as that used by the original storage class in order for mounting of snapshots to succeed. The new storage class should only differ in parameters such as number of replicas.

    CSI snapshot timeout

    The amount of time the agent will wait for a PV snapshot to become ready when mounting. The default is 10 minutes, which should be adequate for most applications. PVs using certain storage systems such as Longhorn may require this to be increased.

  1. In the Destination step, select the destination cluster for the migration. You can select any existing cluster (other than the source) that is registered with CloudCasa, or you can choose to automatically create a new EKS, AKS, or GKE cluster if a cloud account has been registered.

    See also

    If you wish to automatically create a new cluster, see Creating Clusters With CloudCasa.

  2. In the Restore Transforms step, you can choose transformations that will be applied on the destination cluster. These include:

    Rename Namespaces

    Enabling this option will allow you to rename migrated namespaces by either adding a prefix and/or suffix, or applying a mapping.

    Add prefix/postfix

    Add a prefix or postfix to the migrated namespace(s). For example, if there were namespaces “sales” and “services” in the source cluster and you add a suffix, “-dev”, to the namespaces, the migration job will create the namespaces “sales-dev” and “services-dev” in the destination cluster.

    Set new names

    This allows you to create a set of mappings from old namespace names to new namespace names. Names left unmapped will be migrated unchanged.

    Change Storage Classes

    This gives you the option to remap storage classes on the destination cluster. When selected, the UI will display the storage classes used by PVs/PVCs on the source cluster, and allow you to enter new storage classes to substitute for each on the destination cluster. By default, the same storage classes will be used.

    Make sure that a destination storage class is compatible with the source storage class with regards to properties such as supported access modes.

    Preserve node ports

    If selected, automatically assigned node ports for services in the source cluster will be preserved in the destination cluster. Make sure that these ports are available in the destination cluster.

    Note that manually assigned node ports are not affected by this option and are always preserved.

    Overwrite existing resources

    Enabling this option will cause the migration job to overwrite existing Kubernetes resources on the destination cluster. It will do this by attempting to update existing resources to match the corresponding resources from the source cluster. This update is subject to resource update semantics. For example, pod resources don’t allow updates, but most others do. Any resources that cannot be updated will generate a warning in the migration job log.

    Also note that for existing PVCs, the contents of the PVs will not be modified even if this option is selected.

    By default, existing resources will not be overwritten. The exception to this is ServiceAccount resources. Resource data from source cluster ServiceAccounts will be merged into existing destination cluster ServiceAccounts regardless of the overwrite setting.

  3. In the App Hooks step you can choose to add source cluster pre and post-migration application hooks, and destination cluster post-migration application hooks.

    See also

    For more information about App Hooks, see App Hooks.

  4. Review the Summary of the migration job settings to verify that they are correct. In this step you must also enter the following:

    Migration name

    You must assign a name for the migration job.

    By default, new migration jobs will be run immediately. You can toggle off the Run now option if you would rather simply save the job and run it later.

    When you’re done, click Save or Save & Run.

Your new migration job is defined! If you selected the “Run now” option, it will start to execute immediately. If not, you can run it manually later from the Cluster/Migration page or the cluster dashboard.