This is an archive version of the document. To get the most up-to-date information, see the current version.

How Backup Works

In this article

    Veeam Backup for GCP does not install agent software inside VM instances to retrieve data. To back up VM instances, Veeam Backup for GCP uses capabilities of PD snapshots.

    During every backup session, Veeam Backup for GCP creates a cloud-native snapshot of each VM instance added to a backup policy. The cloud-native snapshot is further used to create an image-level backup of the VM instance. Veeam Backup for GCP performs backup in the following way:

    1. Veeam Backup for GCP creates snapshots of persistent disks that are attached to the processed VM instance.

    PD snapshots are assigned resource labels upon creation. Keys and values of resource labels contain encrypted metadata that helps Veeam Backup for GCP identify the related PD snapshots and treat them as a single unit — a cloud-native snapshot.

    1. If you enable image-level backup for the backup policy, Veeam Backup for GCP performs the following operations:
    1. Launches a worker instance in a GCP region where the processed VM instance resides.
    2. Re-creates the persistent disks from the cloud-native snapshot created at step 1 and attaches them to the worker instance.

    Note that the cloud-native snapshot used as a source for image-level backup is not a temporary snapshot — when the backup session completes, this snapshot remains in the snapshot chain and is deleted later according to the specified policy scheduling settings.

    1. Reads data from the persistent disks on the worker instance, transfers the data to a GCP storage bucket and stores it in the native Veeam format.

    Veeam Backup for GCP encrypts and compresses data saved to storage buckets. For more information, see Enabling Data Encryption.

    1. Removes the worker instance when the backup session completes.
    1. If you enable the backup archiving mechanism, Veeam Backup for GCP performs the following operations:
    1. Launches a worker instance in a GCP region where the processed VM instance resides.
    2. Retrieves data from the standard storage bucket and transfers it to the archive storage bucket.
    3. Removes the worker instance when the archive session completes.

    Related Topics