Challenge
This article documents general best practices and configuration advice when using an ExaGrid backup storage appliance with deduplication.Solution
Restores
Restores from recent backups or backup copies will utilize the ExaGrid landing zone as a high-speed disk cache, eliminating the need to rehydrate the backup data. All Veeam vPower operations such as Instant VM Recovery and SureBackup will experience VM boot times of seconds to minutes.
Use the "Per-VM backup files" option on the Storage Compatibility Settings in Repository's "Repository" page (Advanced button). After switching to this option, do a manual active full before the next scheduled backup runs. Set the repository’s “Limit maximum concurrent tasks to:" 10 and work with your ExaGrid support engineer to further tune this setting.
Utilize the ExaGrid Veeam Accelerated Data Mover option whenever possible for best performance and fastest synthetic full and backup copy jobs – since with the Data Mover option, I/O is done directly on the ExaGrid backup storage appliances rather than having to pull Veeam backup data across a network into the proxy server and back out again to the ExaGrid.
You can convert a CIFS repository to use the ExaGrid Veeam Accelerated Data Mover – contact your ExaGrid Customer Support engineer for assistance.
Backup Jobs
- Use “Incremental” backup mode
- Check “Create synthetic full backups periodically” – recommend one day per week
- Check “Enable inline data deduplication” – the ExaGrid will further deduplicate the Veeam backups. See below for more about combined deduplication.
- Change the Compression Level to “Dedupe-friendly”
- Change Storage optimization to “Local”
- If you need to retain more than 14 recovery points, see the section below on how to use Backup Copy Jobs for longer retention
Backup Copy Jobs
When it comes to deciding the best value for number of backup recovery points to retain, keep in mind that the daily incremental backups are, by definition, largely just data that has changed in the VM blocks, and so consume a larger amount of backup storage space since deduplication is not able to significantly reduce the size on disk. ExaGrid standard sizing expects the retention of backup recovery points to be 14 or less. To retain backups for longer than 14 recovery points, utilize Veeam Backup Copy Jobs as follows:
- Consider creating another ExaGrid repository, dedicated for Backup Copy Jobs to ensure that you maintain an isolated copy of your production data to better follow 3-2-1 rule.
- Keep the number of restore points at the minimum of 2, since you will use the backup job’s restore points for your primary source of restore points. Set the copy job interval as long as possible given the GFS settings. So, if GFS settings include weekly retention, set the copy interval to 7 days.
- Create Veeam Backup Copy job(s) to make a copy of the backups you need to retain longer, using the 2nd repository. Enable the "Read the entire restore point from source backup instead of synthesizing it from increments" option on the "Target" tab of the Backup Copy Job.
- Set the retention parameters on the Backup Copy Job to match your needs – i.e. retain 16 weeks, 6 monthlies, 1 yearly, etc.
Using the ExaGrid Veeam Accelerated Data Mover type repository will significantly speed up Backup Copy Jobs, since the reads and write I/O to make the copy all run directly on the ExaGrid backup storage appliance, using the high-speed RAID 6 disks in the ExaGrid appliance rather than having to move all backup data back and forth across between the ExaGrid and the backup proxy across a network.
Contact your ExaGrid Customer Support engineer for assistance in setting up the Veeam Backup Copy jobs for longer retention.
Deduplication
The ExaGrid backup storage appliance will perform deduplication of the Veeam backups, which themselves have been deduplicated by Veeam (see backup job settings above). The ExaGrid web-based administrative interface will show you the ExaGrid-achieved deduplication, but not the Veeam deduplication. For example, if Veeam is deduplicating a backup job 2:1 (typical), and ExaGrid’s GUI show 5:1, the combined dedup is 10:1.
ExaGrid Customer Support has a PowerShell script that collects Veeam deduplication information and produces a combined deduplication result for each Veeam Backup Job; contact your support engineer to obtain the script and assistance using it. Please note that for Veeam Backup and Replication v9, you will need an updated script from ExaGrid Support.