1. Documentation
×

Direct Migration Requirements

Direct Migration is available with Migration Toolkit for Containers (MTC) 1.4.0 or later.

There are two parts of the Direct Migration:

  • Direct Volume Migration

  • Direct Image Migration

Direct Migration enables the migration of persistent volumes and internal images directly from the source cluster to the destination cluster without an intermediary replication repository (object storage).

Prerequisites

  • Expose the internal registries for both clusters (source and destination) involved in the migration for external traffic.

  • Ensure the remote source and destination clusters can communicate using OpenShift Container Platform routes on port 443.

  • Configure the exposed registry route in the source and destination MTC clusters; do this by specifying the spec.exposedRegistryPath field or from the MTC UI.

    • If the destination cluster is the same as the host cluster (where a migration controller exists), there is no need to configure the exposed registry route for that particular MTC cluster.

    • The spec.exposedRegistryPath is required only for Direct Image Migration and not Direct Volume Migration.

  • Ensure the two spec flags in MigPlan custom resource (CR) indirectImageMigration and indirectVolumeMigration are set to false for Direct Migration to be performed. The default value for these flags is false.

The Direct Migration feature of MTC uses the Rsync utility.

Rsync configuration for direct volume migration

Direct Volume Migration (DVM) in MTC uses Rsync to synchronize files between the source and the target persistent volumes (PVs), using a direct connection between the two PVs.

Rsync is a command-line tool that allows you to transfer files and directories to local and remote destinations.

The rsync command used by DVM is optimized for clusters functioning as expected.

The MigrationController CR exposes the following variables to configure rsync_options in Direct Volume Migration:

Variable Type Default value Description

rsync_opt_bwlimit

int

Not set

When set to a positive integer, --bwlimit=<int> option is added to Rsync command.

rsync_opt_archive

bool

true

Sets the --archive option in the Rsync command.

rsync_opt_partial

bool

true

Sets the --partial option in the Rsync command.

rsync_opt_delete

bool

true

Sets the --delete option in the Rsync command.

rsync_opt_hardlinks

bool

true

Sets the --hard-links option is the Rsync command.

rsync_opt_info

string

COPY2

DEL2

REMOVE2

SKIP2

FLIST2

PROGRESS2

STATS2

Enables detailed logging in Rsync Pod.

rsync_opt_extras

string

Empty

Reserved for any other arbitrary options.

  • Setting the options set through the variables above are global for all migrations. The configuration will take effect for all future migrations as soon as the Operator successfully reconciles the MigrationController CR. Any ongoing migration can use the updated settings depending on which step it currently is in. Therefore, it is recommended that the settings be applied before running a migration. The users can always update the settings as needed.

  • Use the rsync_opt_extras variable with caution. Any options passed using this variable are appended to the rsync command, with addition. Ensure you add white spaces when specifying more than one option. Any error in specifying options can lead to a failed migration. However, you can update MigrationController CR as many times as you require for future migrations.

  • Customizing the rsync_opt_info flag can adversely affect the progress reporting capabilities in MTC. However, removing progress reporting can have a performance advantage. This option should only be used when the performance of Rsync operation is observed to be unacceptable.

The default configuration used by DVM is tested in various environments. It is acceptable for most production use cases provided the clusters are healthy and performing well. These configuration variables should be used in case the default settings do not work and the Rsync operation fails.

Resource limit configurations for Rsync pods

The MigrationController CR exposes following variables to configure resource usage requirements and limits on Rsync:

Variable Type Default Description

source_rsync_pod_cpu_limits

string

1

Source rsync pod’s CPU limit

source_rsync_pod_memory_limits

string

1Gi

Source rsync pod’s memory limit

source_rsync_pod_cpu_requests

string

400m

Source rsync pod’s cpu requests

source_rsync_pod_memory_requests

string

1Gi

Source rsync pod’s memory requests

target_rsync_pod_cpu_limits

string

1

Target rsync pod’s cpu limit

target_rsync_pod_cpu_requests

string

400m

Target rsync pod’s cpu requests

target_rsync_pod_memory_limits

string

1Gi

Target rsync pod’s memory limit

target_rsync_pod_memory_requests

string

1Gi

Target rsync pod’s memory requests

Supplemental group configuration for Rsync pods

If Persistent Volume Claims (PVC) are using a shared storage, the access to storage can be configured by adding supplemental groups to Rsync pod definitions in order for the pods to allow access:

Variable Type Default Description

src_supplemental_groups

string

Not Set

Comma separated list of supplemental groups for source Rsync pods

target_supplemental_groups

string

Not Set

Comma separated list of supplemental groups for target Rsync Pods

For example, the MigrationController CR can be updated to set the previous values:

spec:
  src_supplemental_groups: "1000,2000"
  target_supplemental_groups: "2000,3000"

Rsync retry configuration

With MTC 1.4.3 and later, a new ability of retrying a failed Rsync operation is introduced.

By default, the migration controller retries Rsync until all of the data is successfully transferred from the source to the target volume or a specified number of retries is met. The default retry limit is set to 20.

For larger volumes, a limit of 20 retries may not be sufficient.

You can increase the retry limit by using the following variable in the MigrationController CR:

apiVersion: migration.openshift.io/v1alpha1
kind: MigrationController
metadata:
  name: migration-controller
  namespace: openshift-migration
spec:
  [...]
  rsync_backoff_limit: 40

In this example, the retry limit is increased to 40.

Running Rsync as either root or non-root

OpenShift Container Platform environments have the PodSecurityAdmission controller enabled by default. This controller requires cluster administrators to enforce Pod Security Standards by means of namespace labels. All workloads in the cluster are expected to run one of the following Pod Security Standard levels: Privileged, Baseline or Restricted. Every cluster has its own default policy set.

To guarantee successful data transfer in all environments, Migration Toolkit for Containers (MTC) 1.7.5 introduced changes in Rsync pods, including running Rsync pods as non-root user by default. This ensures that data transfer is possible even for workloads that do not necessarily require higher privileges. This change was made because it is best to run workloads with the lowest level of privileges possible.

Manually overriding default non-root operation for data transfer

Although running Rsync pods as non-root user works in most cases, data transfer might fail when you run workloads as root user on the source side. MTC provides two ways to manually override default non-root operation for data transfer:

  • Configure all migrations to run an Rsync pod as root on the destination cluster for all migrations.

  • Run an Rsync pod as root on the destination cluster per migration.

In both cases, you must set the following labels on the source side of any namespaces that are running workloads with higher privileges before migration: enforce, audit, and warn.

To learn more about Pod Security Admission and setting values for labels, see Controlling pod security admission synchronization.

Configuring the MigrationController CR as root or non-root for all migrations

By default, Rsync runs as non-root.

On the destination cluster, you can configure the MigrationController CR to run Rsync as root.

Procedure

  • Configure the MigrationController CR as follows:

    apiVersion: migration.openshift.io/v1alpha1
kind: MigrationController
metadata:
  name: migration-controller
  namespace: openshift-migration
spec:
  [...]
  migration_rsync_privileged: true

    This configuration will apply to all future migrations.

Configuring the MigMigration CR as root or non-root per migration

On the destination cluster, you can configure the MigMigration CR to run Rsync as root or non-root, with the following non-root options:

  • As a specific user ID (UID)

  • As a specific group ID (GID)

Procedure

  1. To run Rsync as root, configure the MigMigration CR according to this example:

    apiVersion: migration.openshift.io/v1alpha1
kind: MigMigration
metadata:
  name: migration-controller
  namespace: openshift-migration
spec:
  [...]
  runAsRoot: true

  2. To run Rsync as a specific User ID (UID) or as a specific Group ID (GID), configure the MigMigration CR according to this example:

    apiVersion: migration.openshift.io/v1alpha1
kind: MigMigration
metadata:
  name: migration-controller
  namespace: openshift-migration
spec:
  [...]
  runAsUser: 10010001
  runAsGroup: 3

MigCluster Configuration

For every MigCluster resource created in MTC, a ConfigMap named migration-cluster-config is created in the Migration Operator’s namespace on the cluster which MigCluster resource represents. 

The migration-cluster-config allows you to configure MigCluster specific values. The Migration Operator manages the migration-cluster-config. 

You can configure every value in the ConfigMap using the variables exposed in the MigrationController CR:

Variable Type Required Description

migration_stage_image_fqin

string

No

Image to use for Stage Pods (applicable only to IndirectVolumeMigration)

migration_registry_image_fqin

string

No

Image to use for Migration Registry

rsync_endpoint_type

string

No

Type of endpoint for data transfer (Route, ClusterIP, NodePort)

rsync_transfer_image_fqin

string

No

Image to use for Rsync Pods (applicable only to DirectVolumeMigration)

migration_rsync_privileged

bool

No

Whether to run Rsync Pods as privileged or not

migration_rsync_super_privileged

bool

No

Whether to run Rsync Pods as super privileged containers (spc_t SELinux context) or not

cluster_subdomain

string

No

Cluster’s subdomain

migration_registry_readiness_timeout

int

No

Readiness timeout (in seconds) for Migration Registry Deployment

migration_registry_liveness_timeout

int

No

Liveness timeout (in seconds) for Migration Registry Deployment

exposed_registry_validation_path

string

No

Subpath to validate exposed registry in a MigCluster (for example /v2)