Grandfathering Space Saver

[Jan2]

Currently in Veeam B&R, I can enable several levels of retention in a grandfathering scheme. The job then retains the weekly full backup for each level for the requested number of copies. For a server with 100GB of data that gets 5 weekly copies and 6 monthly copies retained, the total drive space consumed is slightly in excess of 1 TB. A more manual process for handling daily/weekly/monthly grandfathering exists. It requires a little extra administration. One has to create three jobs: daily, weekly, monthly. Each job has one full and a series of incrementals/reverse incrementals; thereby consuming a bit more than 300 GB. While the process requires more administration, it saves about 800GB of drive space.

There is an opportunity to improve Veeam B&R so that neither of the above processes are necessary, and one can reap substantially more savings in drive space. The key is the reverse incremental process. The most current backup must be a reverse incremental. When retention of increments of a daily backup schedule reach their limit, Veeam B&R should run the increments through the "reverse increment" process and retain a resulting increment and drop all of the daily increments that have expired. This weekly increment becomes the authoritative roll-back point for a restore from a weekly backup. this weekly restore point and subsequent weekly restore points should be retained until the limit of weekly backups. Then, they should be run through the reverse incremental process to create a monthly restore roll-back point. Monthly roll-back points get retained for the limit of their configured retention, and get rolled together through the reverse incremental process to generate an annual roll-back point if configured.

The result for 10 daily, 5 weekly and 6 monthly backups would be: 1 full backup, 6 monthly increments, 5 weekly increments, and 9 daily increments. For a VM that experiences very little change, the total space consumed would be slightly more than 100 GB. For a server that experiences a lot of change, the result should still be less than 300 GB, and requires only one job configuration to manage it.

[HannesK]

Hello,
The recommendation is to go with XFS / REFS and have the benefits of block cloning.

One has to create three jobs: daily, weekly, monthly.

That is more a workaround for everyone who goes with not-recommended file systems

Best regards,
Hannes

[Jan2]

The storage repository has been around longer than Veeam has been using REFS (yeah, it is old). Wiping the entire history of backups just to change drive format is not an option. Perhaps in a few years when I update the storage, I can apply REFS. Thanks for the suggestion.