Reverse Incremental Backup Question

[OhMyEnglishTeaBags]

Hello,

1: If I set a reverse incremental backup to upload to Object Storage, how does the latest full backup get uploaded? Does the incremental merge into the full at Performance Tier and uploads the differential between the two fulls? Or does it have to upload the whole full again?

2: If I set Performance Tier to upload when out of operational window to 0 days, I assume only the full backup and maybe 1 incremental will be stored locally and everything else pushed online?

Here's an example of what i'm trying to achieve: https://ibb.co/RBDRqLG

Thanks in advance.

[Mildur]

Hi Jack

Is this a new infrastructure or an existing one?
I'm asking, if it's a new environment, consider to use a forward incremental chain instead of reverse incremental chain with a reFS or XFS based backup repository.

1: If I set a reverse incremental backup to upload to Object Storage, how does the latest full backup get uploaded? Does the incremental merge into the full at Performance Tier and uploads the differential between the two fulls? Or does it have to upload the whole full again?

We don't store backup files (vbk, vib, vrb) on Object Storage. Only unique blocks of the restore points are offloaded as objects.
The Capacity tier is forever incremental. A Full Backup will only be offloaded the first day, and after that only changed blocks.

2: If I set Performance Tier to upload when out of operational window to 0 days, I assume only the full backup and maybe 1 incremental will be stored locally and everything else pushed online?

The Move Policy can only move inactive backup chains to the capacity tier. In case of a reverse incremental backup chain, there are normally no regular full backups scheduled, so the backup chain is always active. Veeam considers the most recent Full backup and the last two VRB Files as the active backup chain automatically. You will always have this three restore points on your performance tier.
To have all Restore Points on Object Storage from the first day, activate Move and Copy policy for the capacity tier.

[OhMyEnglishTeaBags]

Hi Mildur,

It's a new(ish) environment we configured this week, we've got it set up for Forward Incrementals with Synthetic Fulls. However only now we have realised we can only store 38TB on Performance Tier when we need 80TB. So we were looking for ways to cut corners and Reverse Incrementals seemed an ideal way.

We don't store backup files (vbk, vib, vrb) on Object Storage. Only unique blocks of the restore points are offloaded as objects.
The Capacity tier is forever incremental. A Full Backup will only be offloaded the first day, and after that only changed blocks.

Is this the case for Forward Incremental backups as well then or just Reverse Incremental? Or will Forward Incremental have multiple full backups because of Synthetic Fulls and GFS?

The Move Policy can only move inactive backup chains to the capacity tier. In case of a reverse incremental backup chain, there are normally no regular full backups scheduled, so the backup chain is always active. Veeam considers the most recent Full backup and the last two VRB Files as the active backup chain automatically. You will always have this three restore points on your performance tier.
To have all Restore Points on Object Storage from the first day, activate Move and Copy policy for the capacity tier.

Thank you, this makes sense.

[Mildur]

May I ask, what storage you are using for the performance tier and how much is your source Data (Size of the VMs/Agent)?

Is this the case for Forward Incremental backups as well then or just Reverse Incremental? Or will Forward Incremental have multiple full backups because of Synthetic Fulls and GFS?

This is the case for both backup types. Offload to Capacity tier will always only be changed blocks. Even if you do an active full backup, veeam will only offload unique blocks and not the entire full backup restore point.

[OhMyEnglishTeaBags]

We are using Windows Server 2022 with 38TB Storage. In total, all our VMs are 12,753 GB.

[Mildur]

Do you have local attached disks for the 38TB Storage? If yes, I recommend to format the storage with reFS and use our FastClone technology. You will be able to keep around 10-12 monthly GFS restore points, 4 weeklies and 30 dailies with your storage. Thanks to Fastcloning, which gives you nearly spaceless synthentic full backups. It's only an estimation, it depends on your change rate and how good the data can be reduced.

If I calculate the backup repository requirements for 13TB Source Data, I'll get a storage requirement of 22TB for 30 daily restore points with a weekly synthentic full backup. If you add 12 monthly GFS restore Points, you will require around 38 TB Storage for the backups.
I'll used 50% reduction and a daily Change Rate of 5%:
- https://calculator.veeam.com/vbr/

[OhMyEnglishTeaBags]

Do you have local attached disks for the 38TB Storage? If yes, I recommend to format the storage with reFS and use our FastClone technology. You will be able to keep around 10-12 monthly GFS restore points, 4 weeklies and 30 dailies with your storage. Thanks to Fastcloning, which gives you nearly spaceless synthentic full backups. It's only an estimation, it depends on your change rate and how good the data can be reduced.

If I calculate the backup repository requirements for 13TB Source Data, I'll get a storage requirement of 22TB for 30 daily restore points with a weekly synthentic full backup. If you add 12 monthly GFS restore Points, you will require around 38 TB Storage for the backups.
I'll used 50% reduction and a daily Change Rate of 5%:
- https://calculator.veeam.com/vbr/

Thanks Mildur! I'll re-format to reFS format and give this a go. I wasn't aware this Fastclone existed.

[Mildur]

Your welcome. Let us now how it worked out for you

[OhMyEnglishTeaBags]

Hi Mildur,

I've reformatted to ReFS and set a 64KB cluster but I am unable to find "Use fast cloning on XFS volumes" as outlined: https://www.veeam.com/kb3150

Thanks!

[Mildur]

ReFS is a windows file system.
You will have XFS only on Linux Backup Repository Server. This option is not required for Windows Backup Repository Server.

If you have formated the partiton with ReFS (64KB allocation unit size recommended), veeam will automatically use FastCloning for new Backup Jobs.

[OhMyEnglishTeaBags]

Okay, thank you! I had started to panic!

Many thanks Mildur.

[Mildur]

Your welcome.
Let me know how the job works.
You should see on the first synthetic full in the job details that FastClone was used