Question about GFS Full Backup Size

[tschw_ham]

Hello Veeam Forum,

I administer a Veeam environment. We write data via BackupCopy Jobs to a hardened repository with XFS filesystem. We have created a GFS retention policy that looks like this: 7 Days, 5 Weekly, 12 Monthly.

After about 3 months in operation, the hardened repository was already almost full. We had estimated that with XFS and spaceless fullbackups, there would be room for many more backup sets.

I have entered our data into the Restore Point Simulator (https://rps.dewin.me)

It looks like this:

Style: Backup Copy Job
Used Size GB: 40960
Retention Points: 7
Change Rate: 10%
Data left after reduction: 50%
Interval: Daily
ReFS / XFS: check

Backup Copy Job Specific
Weekly: 5
Monthly: 12
Quarterly: 0
Yearly: 0

Code: Select all

Result
Retention		File		Size	Modify Date		Point Date
-1 -1W 12M 0Q 0Y	full.vbk	20 TB	2021-11-13 Sa 22	2021-11-07 Su 22
-1 -1W 11M 0Q 0Y	full.vbk	10 TB	2021-12-11 Sa 22	2021-12-05 Su 22
-1 -1W 10M 0Q 0Y	full.vbk	10 TB	2022-01-08 Sa 22	2022-01-02 Su 22
-1 -1W 9M 0Q 0Y		full.vbk	10 TB	2022-02-12 Sa 22	2022-02-06 Su 22
-1 -1W 8M 0Q 0Y		full.vbk	10 TB	2022-03-12 Sa 22	2022-03-06 Su 22
-1 -1W 7M 0Q 0Y		full.vbk	10 TB	2022-04-09 Sa 22	2022-04-03 Su 22
-1 -1W 6M 0Q 0Y		full.vbk	10 TB	2022-05-07 Sa 22	2022-05-01 Su 22
-1 -1W 5M 0Q 0Y		full.vbk	10 TB	2022-06-11 Sa 22	2022-06-05 Su 22
-1 -1W 4M 0Q 0Y		full.vbk	10 TB	2022-07-09 Sa 22	2022-07-03 Su 22
-1 -1W 3M 0Q 0Y		full.vbk	10 TB	2022-08-13 Sa 22	2022-08-07 Su 22
-1 -1W 2M 0Q 0Y		full.vbk	10 TB	2022-09-10 Sa 22	2022-09-04 Su 22
-1 5W 0M 0Q 0Y		full.vbk	10 TB	2022-09-24 Sa 22	2022-09-18 Su 22
-1 4W 0M 0Q 0Y		full.vbk	6 TB	2022-10-01 Sa 22	2022-09-25 Su 22
-1 3W 1M 0Q 0Y		full.vbk	6 TB	2022-10-08 Sa 22	2022-10-02 Su 22
-1 2W 0M 0Q 0Y		full.vbk	6 TB	2022-10-15 Sa 22	2022-10-09 Su 22
-1 1W 0M 0Q 0Y		full.vbk	6 TB	2022-10-22 Sa 22	2022-10-16 Su 22
7			full.vbk	6 TB	2022-10-25 Tu 22	2022-10-19 We 22
6			incremental.vib	2 TB	2022-10-20 Th 22	2022-10-20 Th 22
5			incremental.vib	2 TB	2022-10-21 Fr 22	2022-10-21 Fr 22
4			incremental.vib	2 TB	2022-10-22 Sa 22	2022-10-22 Sa 22
3 *W 0M 0Q 0Y		incremental.vib	2 TB	2022-10-23 Su 22	2022-10-23 Su 22
2			incremental.vib	2 TB	2022-10-24 Mo 22	2022-10-24 Mo 22
1			incremental.vib	2 TB	2022-10-25 Tu 22	2022-10-25 Tu 22
							172 TB		
					Work Space	+12.55 TB		
							184.55 TB		

Looking at the Restore Point Simulator result, I wonder why the second Weekly Backup is 10TB in size? I expected it to be only about the size of an increment because it contains only the changes from the previous full backup and the rest of the data is represented via pointer to the previous full backup.

I hope you can help me shed some light on this.

[Mildur]

Changed data between two fulls are not the size of only a daily incremental (2TB).
In this 30 days between the fulls, it‘s possible that most of the data in the vm was changed. So the next full will not take 2TB on disk, it could be much more.

If it‘s only the same data in the vm, which had changed every day, then the full will be much lower (around your expected 2TB).
If different data had be changed through the month, the full will be much higher (veeams calculated 10TB)

Because of that, the calculator is showing you this 10TB.

[tschw_ham]

OK, I think I understand now how this is meant. For the Weekly Backups with 6TB the calculator then simply estimates with less change data?

[foggy]

Yes, and please note that this is just an estimation, in reality, it could be both higher and lower, depending on the data change pattern.

[tschw_ham]

Hello foggy,

thank you for the explanation. This helps me to understand why we have more data stored than we expected.