Running a VM Back On-Prem/On Source
When a VM is moved to run in cloud a base snapshot is taken. When the VM is moved to run back on the source, different behaviors will take place, depending on the storage policy used for that VM.
- Write Isolation: When using write isolation policy, the VM is moved back on-premises to the same point in time captured in the base snapshot. To maintain state alignment across related VMs when they use the write isolation policy, all related VMs should be moved back together.
- Write Back: When running in cloud using the write back policy, and as long as new or changed data is available, a new snapshot (consistency checkpoint) is taken and saved on-premises on every write back cycle (1 hour or 5GB changes, whatever comes first). When you choose to run such a VM on-premises, the remaining data changes are written back and the VM is then moved back on-premises. When the move back is complete, all interim snapshots taken by Velostrata are deleted and the virtual disks are consolidated.
- Forced move back: If there is a sustained cloud outage you may choose to force the VM to run on-premises or on the source. For VMs that use the Write Back policy, this moves the VM back to on-premises using its latest consistency checkpoint. This will result in data loss up to the last stored checkpoint.
Note: As a safety measure, the base snapshot of a VM that runs in cloud using the Write Back policy is not deleted automatically, when the Force option is used. The base snapshot can be deleted manually when verified to be no longer needed.
Note: In the cloud-to-cloud scenario, the source machine is shut down, the storage policy is write isolation, and when moving the VM back to the source, it is returned to the state that it was in when the migration operation started.
Using Web Manager
1. Login to your Velostrata Web Manager at HTTPS://IP_OF_VELO_MANAGER
2. Click the Runbook Automation tab.
3. When prompted for a username and password, use 'apiuser' as the username and your password is your Velostrata subscription ID or your GCP billing ID.
4. Click the monitor icon for one of the runbook automation jobs.
5. Select the VM you wish to return on-premises and click the 'Move Back' button on top.
- On the vSphere Web Client, select the desired Virtual Machine.
- Right-click on the required VM and select Velostrata Operations > Run On-Premises.
- If there has been a prolonged cloud outage, and you want to force a run on-premises, select Force run on-premises.
- Type in the displayed confirmation code to approve the action.
- Click OK. VMs returning from the cloud are automatically started on premises.
The process of running the VM on source can be viewed on the Cloud Instance Information portlet on the VM Summary page, and by monitoring the created vSphere task.
- To move a workload back from GCP to AWS, run the PowerShell command, run: Move-VelosVm [-Id] <String> [-Destination] Origin [Confirm:$false] [-Force]
[Confirm:$false] disables the dialog box warning message.
-Force forces the move back.
For example: Move-VelosVm i-123abc -Destination Origin -confirm:$false