Restoring VMDK Volumes

How to restore volumes is conceptually straightforward. It should be a case of copying the VMDK disks and metadata back to either their original location or to a new location, or restoring the state of an NFS server. However, it is important to understand how vSphere Integrated Containers interacts with the volumes to ensure that this succeeds without error.

• Restoring Volumes into an Existing VCH
• Restoring Volumes into a new VCH

Restoring Volumes into an Existing VCH

Restoring a volume into an existing VCH is a question of copying the VMDK disks and metadata back to the expected location on the datastores.

An important caveat is that you cannot overwrite a volume disk that is attached to a currently running container. If you must bring back a volume that has the same name as one that is in use, there are two solutions:

• Rename the cloned volume
• Rename the cloned volume store.

Renaming a Cloned Volume

Renaming a cloned volume requires 3 steps:

1. Rename the volume folder to newname.
2. Rename the .vmdk file to newname.vmdk
3. Set the value of the name tag in the DockerMetadata file JSON to newname.
4. Copy the volume into the volume store.

5. For vSphere Integrated Containers engine to see the new volume, reboot the VCH endpoint VM.

   You need to reboot because the act of copying the volume into the volume store does not notify vSphere Integrated Containers of its existence. Rebooting the VCH endpoint VM us not unusual or drastic. Any reconfiguration of the VCH causes a reboot of the VCH endpoint VM.

Renaming a Cloned Volume Store

You can bring back an entire cloned volume store under a different name.

1. Rename the root folder of the cloned volume store.
2. Copy it onto a datastore that is visible to the VCH.
3. Run vic-machine configure --volume-store to add the cloned volume store to the VCH.

Restoring Volumes into a new VCH

Restoring a volume store and then creating a new VCH that can use the volumes is a matter of ensuring that the vic-machine create --volume-store argument is correctly configured to point to the volume store. vSphere Integrated Containers is designed to deploy new VCHs onto existing volume stores. You can check whether the volume store has been picked up by running docker info and docker volume ls.