Glacier specific archive tier questions

[theviking84]

1. How will veeam be able to spin up ec2 instances? Are these spun up in my aws account? How is this done if it only was given an archive bucket credential? I haven't tried to set up archive tier so I don't know if it asks for other creds or not.

2. I have not seen an actual step by step guide yet for the process of setting up Glacier archive tier.

3. Do we just make a normal "archive" in glacier in our aws account following aws documentation?

[dalbertson]

1. The IAM account used for the glacier tier will need to have permissions for this. you can see the process of setting up a glacier object repo below in #2.

2. https://helpcenter.veeam.com/docs/backu ... ml?ver=110

3. You just create a "normal" S3 bucket, we use an API to set the storage class of glacier/deep archive for the data.

The IAM permissions needed are listed here https://helpcenter.veeam.com/docs/backu ... ml?ver=110

[theviking84]

Thank you.

[theviking84]

One more glacier question. When adding the bucket and we have the option checkbox for "use deep archive for backups with retention above 180 days" i want to clarify this part.

My expectation is that veeam will use "regular" glacier for all offload to archive tier that are less than 6 month retention in gfs.

Then backups older will go to glacier deep archive.

Meaning BOTH types of glacier are being used with this one setup. Is this the case or no?

Is it instead that anything that goes to archive tier is deep archive if I check it? Doesn't sound that way but just trying to make certain.

So if I have 6 weeklies, 12 monthlies and 2 yearly, then I can expect 6 monthlies and 2 yearlies to go into deep archive while the 6 weeklies and 5 monthlies most recent will go to regular glacier?

[dalbertson]

You were correct. We will use glacier unless you check the box for deep archive older than 180 days. If you setup your GFS to keep everything more than 180 days everything would go to deep archive.

The 180 day mark is because deep archive has an early deletion fee for anything under 180 days.

[theviking84]

Any ETA on when the updated archive tier deep dive will be? I want to try setting it up soon but want more examples/demos and info.

Also, it says in the documentation that up to 64 ec2 proxies can be created during the offloads, and it is per chain I think I read.... So if I have 120 vms it is pretty much going to be 64 proxies and keep spinning them down and up until 120 chains have been done? Isn't this part going to rack up a large bill just for the act of moving the data? Are there any estimates for data size that needs moved or the time it would take to move the data with your proxies to do this?

[Gostev]

Don't even think about this. The cost of moving data to Glacier is negligible comparing to the storage cost. The proxy takes just some cents per hour, while in this hour it will move close to 1TB of data into Glacier. But this 1TB will cost you $12 to store just for the minimal storage duration alone (90 days). So normally there will be 2 orders of magnitude difference between costs of moving and storing.