Disk space requirements for Active Full Backup

[rnt-guy]

Hello,

Is it correct to say that we'd need at least twice the Backup Size in space to keep on iteration in a backup job but do an active full periodically?

So first night say it's 100GB, and second night it's only 10GB but it rolls it in so there's only 1 restore point at the end. Then once a week we do an active full. That would copy over another 100GB, thus making the requirements somewhere around 200GB give or take and then it would delete the previous restore point once it's done?

And if we didn't do active fulls but just did synthetic fulls, it would copy over only that 10GB but still build another 100GB file before deleting the original 100GB?

Am I missing something?

Thanks

[foggy]

What is the backup method (forward/forever forward/reverse) and the number of restore points to keep?

[rnt-guy]

1 restore point.
Currently set to Incremental.

Our goal is to use the least space possible and secondary to take the least amount of time to backup. Restore time is less important in this case.

[foggy]

Are you going to perform active fulls manually? If neither active nor synthetic fulls are scheduled, the job acts in forever forward mode and the increment is merged into existing full backup each time. Once you manually trigger active full, it will require another 100GB of space to be created, but immediately after that the previous full will be deleted according to retention. If any type of full backup is scheduled, the job will act in simple forward incremental mode, where no merge is performed. You will have different number of increments on disk depending on the day of the week.

I will however warn you against having just a single restore point, since with this approach you may fail to have a valid backup at all.

[rnt-guy]

Yes we scheduled them once a week. This way we have 1 week's worth of backups. We don't have the space to keep.

Which takes less space: incremental with scheduled active fulls weekly or reverse incremental with scheduled weekly synthetic fulls?

My understanding is that the active full protects against data corruption the best. From the source perspective the synthetic full takes less time, but overall does a synthetic take less time than an active? The destination disk is probably slower than the source disk.

[csinetops]

Unless you use reverse incremental mode which roll each day into the full backup, you will need enough space to keep 2 fulls on disk. Doesn't matter if you use synthetic full or active full, they both need enough space to create the backup on disk.

There is also forever incremental mode like foggy said where it creates the first active full and then a long chain of incremental after that. While it doesn't use as much space as forward incramental, you have to read up on this mode as I believe restore speeds and risk of losing your chain to a bad backup are greater.

My synthetic fulls take longer than my active, but the advantage is no open VM snaps with synthetic full's putting load on your VM's like with active fulls.All the synthetic full processing is on the backup storage and proxies. I think the key to synthetic fulls is testing your backups periodically to make sure they are good.

https://helpcenter.veeam.com/backup/vsp ... thods.html

[foggy]

If space is your major concern, you should go with either forever forward incremental or reverse incremental: similar space requirements, no periodic fulls, but make sure you perform proper recoverability checks.