Parallel restore for cloud VMs
NetBackup 11.2 introduces parallel restore for cloud VM workloads to reduce restore time for virtual machines with multiple disks. Parallel Restore improves the performance of cloud virtual machine recovery by restoring multiple disks in parallel instead of sequentially.
Supported platforms: AWS and Microsoft Azure
Parallel restore is supported with NetBackup 11.2 or later, subject to the following version requirements:
The datamover (NetBackup Snapshot Manager/DM) must be running version 11.2 or later.
Primary and media servers must be version 11.2 or later.
Note the following:
Parallel restore is supported only when the primary server, media server, and datamover (DM) are all running NetBackup 11.2 or later.
If the primary server is earlier than 11.2, NetBackup performs a sequential restore; parallel restore is not triggered, but the restore completes successfully.
If the primary server is 11.2 or later and the media server is earlier than 11.2, the restore fails. In this case, the media server must be upgraded.
When the primary server, media server, and datamover are all running 11.2 or later, NetBackup performs a parallel restore.
During a parallel restore operation, NetBackup coordinates the restore using a parent job and multiple child jobs. A single parent job controls the overall restore workflow and manages the pre‑restore and post‑restore phases. As part of the pre‑restore phase, the parent job creates the required staging disks for the restore operation.
NetBackup then starts one restore stream per disk. Each restore stream identifies the required data extents for its assigned disk and restores the disk data independently. These restore streams run concurrently, allowing all disks associated with the virtual machine to be restored in parallel. The parent job monitors the progress of the child restore jobs and completes the restore workflow after all disk restores finish successfully and required resources are available.
Parallel restore automatically uses the available media server and storage capabilities to determine, at that time, the number of concurrent restore streams that can run and the number of streams that can be queued. No manual configuration is required to enable this behavior for supported cloud virtual machine workloads.
Parallel restore reduces overall recovery time and improves recovery time objectives (RTOs) for critical workloads.
By default, NetBackup enables parallel restore for supported cloud VM workloads.
Note:
To disable parallel restore add the following parameter to bp.conf file:
CLOUDVM_PARALLEL_RESTORE_DISABLE = 1
Activity monitor will display the parent jobs and child jobs for each stream/disk.
Parallel Restore improves restore performance, particularly for virtual machines with multiple data disks, based on available resources and storage capabilities.