Migration and MigrationSchedule
Synchronous and asynchronous disaster recovery (DR) involves migrating Kubernetes resources from a source cluster to a destination cluster. To set up a migration, you need to pair the source and destination clusters, and then set up a migration schedule to migrate the Kubernetes resources. This document explains the fields that can be used to set up successful migration schedules between the clusters as per your requirements.
MigrationSchedule schema
| Field | Description | Type | Optional | Default |
|---|---|---|---|---|
clusterPair | Specifies the name of the ClusterPair to be used for the migration. | string | No | NA |
adminClusterPair | Specifies the name of the admin ClusterPair used to migrate resources scoped for the cluster, if the ClusterPair is present in a non-admin namespace. | string | Yes | NA |
namespaces | Specifies the list of namespaces to be included in the migration. | string | No Yes (From Stork version 23.5.0) | NA |
namespaceSelectors | When set, resources in the namespaces with the specified namespace labels will be migrated. All the labels provided in this option will be OR'ed. Note that this parameter is only supported on Stork version 23.5.0 and newer. | map [string] string | Yes | NA |
ignoreOwnerReferencesCheck | When set, resources with ownerReferences will also be migrated, even if the corresponding owners are getting migrated. Note that this parameter is only supported on Stork version 23.5.0 and newer. | boolean | Yes | false |
includeResources | Specifies whether Kubernetes resources should be migrated. | boolean | Yes | true |
excludeResourceTypes | Comma-separated list of the specific resource kinds which need to be excluded from migration. For example, Deployment,PersistentVolumeClaim | []string | Yes | NA |
includeVolumes | Specifies whether the underlying Portworx volumes should be migrated. | boolean | Yes | true |
startApplications | Specifies whether the applications should be scaled up on the target cluster after a successful migration. | boolean | Yes | false |
purgeDeletedResources | If Kubernetes resources are removed from the source cluster, this flag specifies to delete them from the target cluster. | boolean | Yes | false |
skipServiceUpdate | Specifies whether service objects should be skipped during migration. | boolean | Yes | false |
includeNetworkPolicyWithCIDR | Specifies whether the underlying network policies should be migrated even if a fixed CIDR is present on them. | boolean | Yes | False |
selectors | When set, only resources with the provided labels will be migrated. All the labels provided in this option will be OR'ed. | map [string] string | Yes | NA |
excludeSelectors | When set, resources with these labels will be excluded from migration. All the labels provided in this option will be OR'ed. Note that this parameter is only supported on Stork version 23.4.0 and newer. | map [string] string | Yes | NA |
| preExecRule | Specifies the name of the rule to be executed before every migration is triggered. | string | Yes | NA |
| postExecRule | Specifies the name of the rule to be executed after every migration is triggered. | string | Yes | NA |
| includeOptionalResourceTypes | Set this flag to ensure that the Kubernetes resources associated with [Job] (https://kubernetes.io/docs/concepts/workloads/controllers/job/) are migrated. By default, the Job resources are not migrated. | []string | Yes | NA |
skipDeletedNamespaces | Specifies if migration should ignore deleted namespaces during migration. If set to false, Stork will fail the migration when it encounters a namespace that is deleted but specified in the namespaces field. | boolean | Yes | true |
transformSpecs | Specifies the ResourceTransformation spec to be applied during migration. | []string | Yes | NA |
note
You must specify either the namespaces or the namespaceSelectors parameter.