Repository Mainenance

[nunciate]

I'd like to do some kind of feature request or ask if anyone can tell me how they manage their repository utilization.

I have 4 drives running on a Windows 2019 server (REFS). They are all 46Tb each. I have them all in a Scale-Out repository.
I am constantly managing the drive space utilization. I want to save 30 restore points per backup but I either fill up all the drives or I fill up 1 drive which kills a bunch of my backup jobs.

I slowly reduced the number of restore points down to 20 but I still fill up some drives because the backup jobs put too many active full backups on the same partition.

The request I have is to be able to limit the drive use when it becomes say 80% full. Also, something needs to be done to determine better placement for vbk files as putting a lot on one drive is bad.
I get a full drive and then when it comes time to rollup changes into a full the jobs fail. It is quite frustrating and takes up a lot of my time to manage this.

As I said above I am using refs so I cannot use Windows quotas to manage that and I don't want to reformat my repositories just to switch to NTFS. The drives use Windows deduplication set to the general file server option with a dedupe after 3 days.
Each repository is set to align backup file data blocks, decompress backup file data blocks before storing and user per-machine backup files.
The scale-out repository is set to use per-machine backups also and the placement policy is set to performance.

[Mildur]

Hi Alan

I suggest changing your settings if you want to have less issues with space.
- disable reFS deduplication
Our recommendation is to use our FastClone. Spaceless and fast Synthetic fulls. FastClone cannot be used when you enable reFS deduplication.

- switch to placement policy data locality
Full Backup and incremental backups are on the same disk. FastClone will allow you to reuse the blocks of a synthetic backup for full backup files.

- run only synthetic fulls.
FastClone aware backups.

I expect less space issues with our recommended storage solution. And better backup performance with synthetic fulls. Synthetic fulls are target to the same extend as the last full and incremental of the same job.
Active full backups can go to any extend and fill your disk.

Best,
Fabian

[nunciate]

Thanks for the info. I turned off deduplication on the volumes. I plan to remove the existing backups from disk and re-run the active full on each to get the old dedpulcated data off the volumes.
I'd keep the existing backups but I don't have the space right now.