Archive Glacier Different Acconts

[sarapinho]

I have two buckets set up on my Veeam with different accounts.

Account 01 is configured as S3 (Capacity Tier).
Account 02 is configured the S3 (Glacier)

Will it be possible for me to archive to Glacier the buckets being in different accounts?

The proxies are in the same bucket account as Glacier.

[Gostev]

Yes, should be possible... I am not aware of any limitations here. We repackage archived data into much bigger objects for cost efficiency, and should be able to write the resulting objects to any bucket.

[sarapinho]

Hi Gostev.

Thanks for the answer,

So I need to release my S3 (Capacity Tier) so that the proxies that are on the S3 (Glacier) account have access, correct? And this access will be through the WAN ip?

[Gostev]

It should just work as Veeam has credentials to both buckets, you shouldn't need to do anything special with access. Don't overthink this, at least until you actually have issues with something not working

[sarapinho]

Ok, but in this scenario the archive is not working.

Case # 05061122

[veremin]

We will double check the case internally and come to you later this week.

[sarapinho]

Thank you for your attention

[veremin]

The case is in Portuguese so I cannot make a sense of it.

Generally, this configuration should work as long as you:

* Configure Cross-Account access to the S3 bucket (option 3 from this article)
* While adding AWS S3 Glacier repository specify bucket name manually instead of browsing for it

Out of curiosity - are there are any specific requirements or use cases that make you think about Cross-Account scenario?

Thanks!

[sarapinho]

Ok, no problem regarding the case being Portuguese!

Yesterday, we performed several together with support (Thanks Rafael Oliveira), and we discovered an access point configuration that was not configured. After changing the access point the archive worked.

Let's redo every part of acl in aws limiting only to our ip of wan and then post the result.

Thanks for the support!

[veremin]

Great news, Amauri, thanks for sharing - looking forward to the final results.

[sarapinho]

Hello everybody,

Sorry for the delay in replying, the inter-account archive didn't really work! What was informed by aws support would be a large data movement and also cost, which was discarded by the customer.

We enabled archive on the same account, and it worked. But a doubt arose!

During the archive process my S3 bucket also had a capacity increase, is this normal?

I did a 10TB archive and my bucket also expanded 10TB

[Gostev]

Yes, it's normal.

Those new objects in your S3 bucked have a different object class and are thus a few times cheaper to store.
See my sample archiving cost efficiency calculation here (this is for Azure, but same idea).

[sarapinho]

I don't think I understand... If I'm moving to the archive tier, shouldn't my capacity decrease?

Example: I have 300Tb in the capacity tier, I will move archive backups older than 180 days. Let's assume 150TB is going to be moved to the archive, ok?

It shouldn't stay in this format:

150Tb in archive and 150Tb in capacity?

[Gostev]

Your Capacity Tier (CT) will decrease, but only by relatively little because you're still storing 180 days worth of backups there. Remember, only blocks that were exclusive to archived restore points can be removed from CT. Other blocks cannot be removed, because they are a part of those 180 days worth of restore restore points, which are still stored in Capacity Tier.

It's impossible to say what will be the split between CT and AT because it depends heavily on a change rate, as well as how many years of backups end up in AT after the move.

If you want to significantly reduce CT storage consumption, then you should look at reducing your archiving window from 180 days significantly, for example to 2 weeks.