Configuring VM-Centric Snapshots for the Tintri VMstore Storage Array

When you perform IntelliSnap backups with the Tintri storage array, the backup takes a hardware snapshot of each virtual machine being backed up, instead of taking snapshots of the entire datastores that host VMs.

This feature is supported for VMware IntelliSnap backups.

Before You Begin

  • On the ESXi server that will be used to mount VMs for backup copies and restores, create an NFS datastore (for example, VMCentricBackup).

  • Perform the following configuration in the Tintri management portal:

    • In VMstore hypervisor manager settings, add the fully qualified domain name for the vCenter and provide the credentials for the vCenter.

    • Datastore hosted as a Tintri NFS Share (for example, /tintri/VMCentricBackup).

Procedure

  1. In the Array Management settings for the Tintri array, add the IP address for the Tintri array.

  2. In the Array Properties for the Tintri array, on the Snap Configuration tab, add the IP address for the array datapath for the Array Host Aliases settings.

    This step is not required if the IP address for data is the same as the IP address for management.

  3. Identify a Tintri datastore that can be used as the mount path. A single mount path can be used for all jobs for the vCenter.

  4. Configure the subclient for virtual machines to be backed up to the Tintri array.

  5. On the Backup Options tab of the Subclient Properties dialog box, select one of the following settings:

    • File System and Application Consistent: Use VMware tools to quiesce the operating system and applications on each virtual machine in the subclient during the backup of the VM. The backup process may slow down because of the quiescing. VMware Tools are required to quiesce the operating system and applications.

    • Crash Consistent: Perform a backup without quiescing the operating system and applications.

  6. Perform a backup for the subclient.

IntelliSnap Backup and Restore Process

  1. For each VM in the subclient, the backup operation creates a separate archive file.

  2. The backup uses a separate stream for each VM.

    The number of streams that are available for a backup job depends on the number of data readers that are configured for the subclient.

    Backups are performed asynchronously in parallel.

  3. The backup creates a catalog file for each VM and updates the index file for the backup job.

  4. VM snapshots are retained on the array.

  5. For backup copy and restore operations, the VM snapshot is used to create a smart clone VM on the NFS datastore configured for the array.

  6. The cloned VM is registered on the ESXi server and is powered off.

  7. The cloned VM is used as a source for the backup copy or restore operation.

  8. When the backup copy or restore operation is completed, smart clone VMs are unregistered from the ESXi server and then deleted.

Known Issues

  • Live browse, live file recovery, live mount, and live VM recovery are not supported.

  • You cannot mount or unmount the snapshots through the CommCell Console.

  • VMs that span different VMstores (Tintri arrays) cannot be backed up.

  • VMs that have multiple disks hosted on different datastores from the same VMstore cannot be backed up.

  • A backup using VM-centric snapshots does not collect file and folder metadata.

  • Incremental backups use cyclic redundancy checking (CRC) instead of changed block tracking (CBT), because the Tintri framework does not support CBT.

Loading...