Procedure for performing Snaps using Solutions Enabler TimeFinder Snapvx on VMware VMDK devices

Virtual Machine shuts down when attempting to bring drives online on Snap target.

More procedure is required when performing snapvx link on VMware VMDK drives.

The following is the recommended procedure when running TimeFinder SnapVX on VMDK drives.

Datastore refresh

• In test environments, it may be desirable to refresh the linked target with the original snapshot at some point. The refresh procedure is slightly different depending on whether the linked target device was resignatured or not. Both processes are included below.

Resignature refresh
When the linked target is resignatured, as recommended, and then refreshed, the new VMFS signature is overwritten. Therefore it is essential that the datastore is removed, and the linked target is not in use during the relink procedure. To ensure a clean refresh, the following steps should be used.

1. Shut down all VMs located on the linked target datastore

2. Remove the VMs from inventory

3. Delete the datastore

4. Detached the device (optional, though recommended)

5. Relink the target device

6. Repeat the process for detaching and adding back the linked target

Keep signature refresh
When the linked target is mounted with the same signature, a refresh of the linked target will not overwrite it, and therefore the datastore does not have to be deleted. There are two different procedures that can be used. The first one ensures a clean refresh but takes longer. The second is quicker but has the potential to cause issues. Either is acceptable. The following are the recommended steps for a clean refresh.

7. Shut down all VMs located on the linked target datastore

8. Remove the VMs from inventory

9. Unmount the datastore from each ESXi host using CLI

10. Relink the target device

11. Re-mount the datastore on an ESXi host using CLI

12. Re-register the VMs

The following steps provide a quicker refresh.

13. Shut down all VMs located on the linked target datastore

14. Relink the target device

15. Refresh the contents (within the vSphere Client)

The following are some caveats around this procedure.

• If the datastore fails to refresh (folder dates do not revert), or simply shows no content, it may be necessary to unmount and remount the datastore on each ESXi host with the CLI.

• If the registered VMs from the datastore are named differently than the original VMs, for example, the VM in the linked target was named TEST, but instead the VM was registered as TEST_VM, the registered VM reverts to the original VM name. This may occur automatically, or when a power-on is attempted (the power-on will error and at that point the VM will rename).

It is possible some other unforeseen event may occur; hence the clean refresh is recommended.