Stay organized with collections
Save and categorize content based on your preferences.
My workload does not start
When attempting to start a migration, you may encounter an error that
prevents your workload from starting correctly.
If you experience an error that prevents your workload from starting correctly,
try the troubleshooting steps described in this document before contacting
support.
Add permissions required to pull images from Google Container Registry
For your workload to start, the cluster needs to pull the workload image from
Google Container Registry (GCR), which might sometimes
fail due to missing permissions.
To identify this issue, perform these steps:
In the Google Cloud console, go to the Object browser page.
From the list of pods that is displayed, locate the pod corresponding to your
workload, and then to open pod details, click the pod name.
On the Pod details page, if a banner appears that displays the
failed to pull and unpack image and 403 forbidden errors,
then the permissions required to pull the workload image are missing.
To resolve this issue, perform these steps:
Add a role
called Storage Object Viewer to the Default Compute Engine service account
in your project.
A new pod is automatically created which replaces the deleted pod.
Your migrated workload should now be accessible.
Disable GKE Autopilot clusters
As of Migrate to Containers, use of GKE Autopilot clusters is
enabled by default. As a result, any new migrations created for
Migrate to Containers will use GKE Autopilot clusters unless specified
otherwise.
Try disabling GKE Autopilot clusters and attempt to start
your migration workload again.
To disable GKE Autopilot clusters, perform these steps to set
v2kServiceManager to false:
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-08-28 UTC."],[],[],null,["# My workload does not start\n==========================\n\nWhen attempting to start a migration, you may encounter an error that\nprevents your workload from starting correctly.\n\nIf you experience an error that prevents your workload from starting correctly,\ntry the troubleshooting steps described in this document before contacting\nsupport.\n\nAdd permissions required to pull images from Google Container Registry\n----------------------------------------------------------------------\n\nFor your workload to start, the cluster needs to pull the workload image from\n[Google Container Registry (GCR)](/container-registry), which might sometimes\nfail due to missing permissions.\n\nTo identify this issue, perform these steps:\n\n1. In the Google Cloud console, go to the **Object browser** page.\n\n [Go to Object browser](https://console.cloud.google.com/kubernetes/object/browser)\n\n \u003cbr /\u003e\n\n2. Select your cluster.\n\n3. From the **Object Kinds** list, select **Pod**.\n\n4. From the list of pods that is displayed, locate the pod corresponding to your\n workload, and then to open pod details, click the pod name.\n\n5. On the **Pod details** page, if a banner appears that displays the\n `failed to pull and unpack image` and `403 forbidden` errors,\n then the permissions required to pull the workload image are missing.\n\nTo resolve this issue, perform these steps:\n\n1. [Add a role](/iam/docs/manage-access-service-accounts#grant-single-role)\n called **Storage Object Viewer** to the **Default Compute Engine service account**\n in your project.\n\n2. Then, [delete](https://kubernetes.io/docs/reference/generated/kubectl/kubectl-commands#delete)\n the pod from your cluster.\n\n A new pod is automatically created which replaces the deleted pod.\n\nYour migrated workload should now be accessible.\n\nDisable GKE Autopilot clusters\n------------------------------\n\nAs of Migrate to Containers, use of GKE Autopilot clusters is\nenabled by default. As a result, any new migrations created for\nMigrate to Containers will use GKE Autopilot clusters unless specified\notherwise.\n\nTry disabling GKE Autopilot clusters and attempt to start\nyour migration workload again.\n\nTo disable GKE Autopilot clusters, perform these steps to set\n`v2kServiceManager` to `false`:\n\n1. [Edit your migration plan](/migrate/containers/docs/customizing-a-migration-plan#edit_the_migration_plan).\n\n 1. In the \u003cvar translate=\"no\"\u003eMIGRATION_NAME\u003c/var\u003e`.yaml` file, locate\n `v2kServiceManager` and set it to `false`.\n\n Change: \n\n v2kServiceManager: true\n\n to: \n\n v2kServiceManager: false\n\n 2. Save your file.\n\n2. [Reinitiate your migration](/migrate/anthos/docs/executing-a-migration) using Migrate to Containers.\n\nIf your workload still does not start correctly after disabling GKE Autopilot clusters,\nthen contact your support channel."]]
