Scale out repository - add existing repo

[jja]

Hi!

I just upgraded our B&R to V9, and I'd like to start using scale out repositores.
We only have 7TB free space on our repo, and I have a new server soon to be added.

If I create a new scale out repo and add the existing repo with 80+TB of backups, will everything still be available as before?
I'm pretty sure it's possible, but it's better to ask and be safe

-Jannis

[dellock6]

Hi Jannis,
yes nothing is touched on the existing repositories, existing backup files are only imported into the scale-out repo and related jobs are remapped to the new destination.
Only be careful, space is not automatically rebalanced even when you add the new machine, only new active fulls (or new vm as scale-out will switch to per-vm chains) will start to consume the extent with the bigger free space.

Luca

[jja]

So all my jobs going to the existing repo which is nearly full, will still go to the same repo regardless of free space on other extents?
Is there a simple way to get the jobs to start using the new extents without loosing anything, and not having to reconfigure a lot?

-Jannis

[veremin]

As suggested, you can initiate a new active full backup cycle for existing jobs in order for them to start consuming newly-added extent. Thanks.

[jja]

Will this do a "full" full backup (eg. full read and store of vmdks)?
If so, we would loose a lot of the space on the new storage

-Jannis

[mkretzer]

What about when the other extend gets full? Will it re-balance the data after the first active full (for example if a synthetic is created)?

[veremin]

Depends on amount of free space available on a new extent and backup placement policy. Other factors are mentioned here. Thanks.

[jja]

Pretty much all our jobs are reverse incremental.
And it seems that it automatically started using the new extent.
Is this because the new backup always is the full backup?
either way, this saved us for a lot of hassle
(even though the reverse incremental has a higher IO cost)

Is it recommended to change to an incremental policy instead?

-j

[veremin]

Is this because the new backup always is the full backup?

There is a dedicated section of User Guide describing extent selection logic that I've referenced above. So, kindly, take a look.

Is it recommended to change to an incremental policy instead?

If current situation suites your needs, you can leave everything as is.

Thanks.