Migrating a Standard Virtual Machine using OpenShift Migration Toolkit for Virtualization

This is the baseline workflow for migrating a VM with one or more disks that are not shared with other VMs.

1. Ensure that xcopy offload is enabled with MTV

   oc patch forkliftcontrollers.forklift.konveyor.io forklift-controller \
     --type merge \
     -p '{"spec": {"feature_copy_offload": "true"}}' \
     -n openshift-mtv
   

2. Create the Forklift Project (or use an existing project if desired, in the GUI example the openshift-mtv project is used)

   oc new-project migration-project

3. Define the Source Provider:

   • Using the OCP GUI

     Navigate to Migration for Virtualization -> Providers. Then click on "Create Provider".

     

     As we will be leveraging xcopy or vol copy offload to the FlashArray, we did not configure the VDDK.

     

     Provider the provider credentials and then configure the certificate validation or choose to skip.

     

     Review the provider to ensure that everything looks correct.

     

   • Using CLI:

     # Create vCenter credentials secret
     oc create secret generic "vcenter-source-creds" \
     --from-literal=user="user_name@vsphere.local" \
     --from-literal=password="super_secret_password" \
     --from-literal=insecureSkipVerify=true \
     -n "migration-project" --dry-run=client -o yaml | oc apply -f -
     # Define the Provider CR
     # Note: vddkInitImage is omitted — not required for XCopy Offload
     cat << EOF | oc apply -f -
     apiVersion: forklift.konveyor.io/v1beta1
     kind: Provider
     metadata:
       name: vcenter-source
       namespace: migration-project
     spec:
       type: vsphere
       url: https://<vcenter-ip-or-fqdn>
       secret:
         name: vcenter-source-creds
         namespace: migration-project
     EOF
     

4. Define Network and Storage Maps:

   • Using the OCP GUI

     Create the FlashArray Storage Credential Secret

     First you need to retrieve the prefix for the px cluster. In this example a Pod is being used for the destination. From the FlashArray GUI, navigate to Storage => Pods => Select the Pod => Find a volume in the pod.

     

     Looking at the volume, here is the prefix that will be used in the storage provider credentials to leverage xcopy and vol copy offload.

     

     Next the secret needs to be created. Navigate to the Workloads => Secrets and then create from yaml

     

     Use the following YAML with the environment for the FlashArray and vCenter

     apiVersion: v1
     kind: Secret
     metadata:
       name: pure-storage-creds
       namespace: openshift-mtv
     Data:
       GOVMOMI_HOSTNAME: prod-vcsa.alex.purestorage.com
       GOVMOMI_USERNAME: carver@alex.purestorage.com
       GOVMOMI_PASSWORD: super-secret-password
       PURE_CLUSTER_PREFIX: "openshift-03-b24-29-local-pod-01::px_90f5c7e8"
       STORAGE_HOSTNAME: 10.21.147.60
       STORAGE_USERNAME: pureuser
       STORAGE_PASSWORD: super-secret-password
       STORAGE_SKIP_SSL_VERIFICATION: true
     type: Opaque
     

     Here is an example of the secret being created.

     

     Create the Storage Map

     From Migration for Virtualization => Storage Maps => Create storage map - choose to create with the form.

     

     The form outlines the required information for creating the storage map. Use the vSphere xcopy offload, the secret that was just created and then select the Pure FlashArray.

     

     Create the Network Map

     From Migration for Virtualization => Network Maps => Create network map - select to create from form.

     

     Provide the required information and ensure that the source network maps correct to the network on the subnet for the VM.

     

   • Using CLI and YAML:

     # FlashArray Secret Storage Map — creates credentials for FlashArray for migration
     oc create secret generic pure-storage-creds \
     --from-literal=STORAGE_HOSTNAME="<redacted-storage-endpoint>" \
     --from-literal=STORAGE_USERNAME="pureuser" \
     --from-literal=STORAGE_PASSWORD="<redacted-password>" \
     --from-literal=GOVMOMI_HOSTNAME="<redacted-hostname>" \
     --from-literal=GOVMOMI_USERNAME="<redacted-username>" \
     --from-literal=GOVMOMI_PASSWORD="<redacted-password>" \
     -from-literal=PURE_CLUSTER_PREFIX="openshift-03-b24-29-local-pod-01::px_90f5c7e8
     --from-literal=STORAGE_SKIP_SSL_VERIFICATION="true" \
     -n "openshift-mtv" --dry-run=client -o yaml | oc apply -f -
     # NetworkMap — maps VMware port group to OpenShift pod network
     apiVersion: forklift.konveyor.io/v1beta1
     kind: NetworkMap
     metadata:
     name: default-network-map
     spec:
     map:
     - destination:
     name: pod
     type: pod
     source:
     id: <vm-network-id> # Retrieve from vCenter API
     name: <vm-network-name>
     provider:
     source:
     name: vcenter-source
     namespace: ${FORKLIFT_NAMESPACE}
     destination:
     name: host
     namespace: openshift-mtv
     # StorageMap — maps VMware datastore to Portworx + FlashArray StorageClass
     # offloadPlugin section enables XCopy Offload
     apiVersion: forklift.konveyor.io/v1beta1
     kind: StorageMap
     metadata:
     name: pure-storage-map
     spec:
     map:
     - destination:
     accessMode: ReadWriteOnce
     storageClass: px-pure-iscsi-sc # or fc StorageClass
     offloadPlugin:
     vsphereXcopyConfig:
     secretRef: pure-storage-creds
     storageVendorProduct: pureFlashArray
     source:
     id: <source-datastore-id>
     name: <source-datastore-name>
     provider:
     source:
     name: vcenter-source
     namespace: ${FORKLIFT_NAMESPACE}
     destination:
     name: host
     namespace: openshift-mtv
     # Apply both maps
     oc apply -f network-map.yaml -n ${FORKLIFT_NAMESPACE}
     oc apply -f storage-map.yaml -n ${FORKLIFT_NAMESPACE}

5. Define the Migration Plan:

   • Using the OCP GUI

     From Migration for Virtualization => Migration Plans => Create Plan

     

     Provide the plan name, project, source provider, target provider and target project.

     

     Provide the network map

     

     Provide the storage map

     

     Select Cold Migration - Everpure has done testing with Warm vs Cold and we have validated that warm migration works, but it will require the VDDK to be configured and at the time of this article there was a longer migration time and VM downtime during warm migration vs cold migration with storage offload. At this time the recommended migration type is Cold Migration.

     

     From the other settings, select the ones that apply for the VM

     

     Review the plan details

     

   • Using CLI:

     apiVersion: forklift.konveyor.io/v1beta1
     kind: Plan
     metadata:
     name: standard-vm-plan
     spec:
     pvcNameTemplateUseGenerateName: true
     targetNamespace: ${FORKLIFT_NAMESPACE}
     warm: false
     provider:
     source:
     name: vcenter-source
     namespace: ${FORKLIFT_NAMESPACE}
     destination:
     name: host
     namespace: openshift-mtv
     map:
     network:
     name: default-network-map
     namespace: ${FORKLIFT_NAMESPACE}
     storage:
     name: pure-storage-map
     namespace: ${FORKLIFT_NAMESPACE}
     vms:
     - id: vm-1001 # VMware managed object reference ID
     name: my-test-vm

6. Execute and Monitor:

   • From the OCP GUI

     To execute the migration select "Start" this can be done from the migration plan main view or from clicking on the migration plan itself.

     

     Once the migration has started you can monitor the migration process from the Virtual Machines view on the migration plan details.

     

     Once the migration has completed you can review or debug each of the steps.

     

   • From the CLI

     # Create Migration CR to start execution
     apiVersion: forklift.konveyor.io/v1beta1
     kind: Migration
     metadata:
     name: standard-vm-migration
     spec:
     plan:
     name: standard-vm-plan
     namespace: ${FORKLIFT_NAMESPACE}
     # Apply the plan
     oc apply -f migration-plan.yaml -n ${FORKLIFT_NAMESPACE}
     # Monitor migration status
     oc get migration standard-vm-migration -n ${FORKLIFT_NAMESPACE} \
     -o jsonpath='{.status.conditions[?(@.type=="Succeeded")].status}'

     Note: For a Windows 10 VM with 100 GB disk (50 GB used): XCopy Offload completes in 4m 07s. VDDK takes 15m 07s. No difference between disk provisioning modes.