How to configure DFSR backups to use Accelerator with NetBackup versions

Article: 100012173
Last Published: 2020-03-23
Ratings: 1 1
Product(s): NetBackup & Alta Data Protection

Problem

DFSR (Distributed File System Replication) data sets can be very large.  Accelerator is designed to improve backup performance of very large data sets.

 

Solution

Accelerator notes and requirements

The key steps are the following:

  • The Master Server, Media Server, and Client should all be 8.x and newer
  • Implement a NetBackup Accelerator License Key on the Master Server
  • Configure the DFSR Policy with an Accelerator supported Storage Unit (MSDP, PureDisk, etc...)
  • Configure the DFSR Policy by checking the "Use Accelerator" checkbox
  • In addition to the regular weekly/monthly/yearly FULL schedules, configure an additional FULL schedule which has "Accelerator Forced Rescan" enabled
    NOTE: This schedule should be configured to run on a seldom, yet regular basis depending on your business needs, typically about 1 time every 3-6 months. 
  • Please review the  Accelerator forced rescan option (schedule attribute) section in the Veritas NetBackup Administrator's Guide, Vol I.
  • Policy Backup Selections must specify each replicated folder individually. See article 100014007 for details  

Anticipated observed behavior:

  • The first Accelerator enabled FULL backup will need to build the Accelerator Track Log - this job will run with the same performance (or maybe a little worse) then a non-Accelerator enabled FULL
  • The second and ongoing Accelerator enabled FULL backups will observe the performance benefit Accelerator is designed to deliver
  • The "Accelerator Forced Rescan" schedule will rebuild the Accelerator Track Log and will run with the same performance as the initial Accelerator enabled FULL

These conditions will eliminate the performance enhancement Accelerator is known to deliver:

  • If an "Accelerator Forced Rescan" schedule does not exist in the DFSR Policy, the regular FULL schedule policies will always perform a rescan and rebuild the track log each time.  
  • If the Client is at 7.x , and the Master Server and/or Media Servers are at 8.x , the regular weekly/monthly/yearly FULL schedules will also force a rescan and rebuild the track log each time.

 

 

Was this content helpful?