This section addresses the workflow for migrating an application. For a view centered on the VM, see the VM migration lifecycle.
Migration planning occurs in multiple stages:
Modeling and planning migration waves: Before beginning a migration, first create a prioritized inventory of the machines you want to migrate. You also need to understand your application's dependencies and use Migrate for Compute Engine's Wave interface to create migration waves and plan the resulting migrations. Then, you need to batch these dependencies within the same wave. For example:
- If an application depends on multiple VMs, migrate all of those VMs in the same wave.
- Alternatively, if an application requires a database and a web server, the database can be migrated and started before the web server.
Pre-migration testing (optional). To test your migration before running it, you can clone an on-premises workload and run it in Google Cloud without modifying the original host. This enables you to test without disrupting production. Clones can be fully operational within minutes. For more information, see Creating a test clone.
Cutover to cloud. Run the VM in Compute Engine in streaming mode. This can be done per VM or using Waves for mass migration sprints. You can accomplish the following phases as part of the migration to cloud, using the Migration Wizard or using Waves.
Move the VM to Compute Engine.
See Running and migrating VMs. After this is completed, perform validation or sanity tests and fix the resulting issues.
Migrate the application storage to Google Cloud.
Detach. After determining that you can rely on the VMs in the cloud, schedule downtime to cut over your application to Google Cloud. This happens in two stages:
- Detach the VM. Follow these instructions to detach a VM.
- Test the application to verify that it functions accurately post-migration.