DR - testing the "scorched earth" scenario using S3 and AWS

[hoffmp]

Hi all

I could use a little help getting past what I imagine is a simple stumbling block. We back up our on premises VMware environment to local storage and to a S.O.B.R. in which Amazon S3 acts as the capacity tier. Our "scorched earth" scenario assumes that all local hardware and storage is gone. (Assume the building burnt to the ground) The plan is to
1. Create a Windows EC2 instance in AWS
2. Install Veeam BU&R on that instance
3. Connect to backups as stored in S3 via the S.O.B.R.
4. Restore what were VMs to Amazon EC2 instances

I'm getting stuck at step 3. We've had zero problems doing test restores from S.O.B.R. back to the on-premises VMware environment. But I get errors just connecting to and rescanning the S3 capacity tier. And worse yet, I don't see all of the VMs represented in the inventory. I suspect I'm skipping an obvious step, but I haven't been able to find a process for connecting a new Veeam BU&R server to an existing S.O.B.R. repository. If you can point me towards such a document or outline the procedure, it would be much appreciated. Thanks

Almost forgot... we're running Veeam BU&R version 10

Phil

[dalbertson]

Have you tried importing the backups?

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

[hoffmp]

I may have been heading down the wrong path and trying to recreate the S.O.B.R. If I understand correctly, I'd only have to recreate an Object Repository pointing to the S3 and then import?

[dalbertson]

Correct. You don’t want to create a new sobr.

Just create the object repo and import the backups and start restoring away

[hoffmp]

Thanks. Much easier than what I was attempting. Will give it a try and post an update

[hoffmp]

Sorry, but I'm still having trouble. On my "in house" Veeam server, everything looks great. But when I try to create the External Repository for S3, I can see the bucket but none of the folders within the bucket

[dalbertson]

Ok there lies the problem.

External repo is not for use like that. It’s used with our cloud backup products. This allows vbr to see the backups created by those products.

You should be deploying a new vbr server in aws and then creating a object repository under backup infrastructure - backup repositories. Once you connect to the same S3 bucket it will ask if you want to import the existing backups.

[hoffmp]

OK. First and foremost, I sincerely appreciate your help and patience. Thank you.

I've attached the repository successfully. Seemed to connect to the new BU&R server correctly, but I don't have any way to import the backups. The option shown in the link you were kind enough to provide earlier doesn't appear to be available

[dalbertson]

No worries. Is the “dr” vbr server licensed? Also can you send a screenshot of the repo when you right click on it

[hoffmp]

Yes. I applied the same license that's used on the in-house server, assuming that would be ok since we're not protecting any new environments.

Sorry, another dumb newbie question -- not sure how to send you an image using this forum

[dalbertson]

You can use any image hosting service and then just paste the link in here

Like https://imgbb.com

[hoffmp]

Got it. https://photos.app.goo.gl/eTZSH3zntcfBUQg49

[dalbertson]

Thank you. Can you show me your license screen too?

[hoffmp]

Certainly. Sorry for obfuscating some of the info, but as this is a public forum, I think it prudent to do so
https://photos.app.goo.gl/qWJMuv6XKPDMxE499

[veremin]

Can you also show how the said repository is shown in backup console and what actions are available, if you right-click on it? Thanks!

[hoffmp]

All I can see in the console is this: https://photos.app.goo.gl/eTZSH3zntcfBUQg49

[veremin]

It seems that backups from this repository have been imported successfully. Can you go to Home -> Backups and check "Object Storage (Imported)" node? Can you see the imported backups there? Thanks!

[hoffmp]

And that's when it gets a bit strange. What I see is this https://photos.app.goo.gl/SBMu4xykZwp2dy1i7 But there are more jobs and many more restore points visible from the on-premises BU&R server that writes to this S.O.B.R.

[veremin]

This is strange, indeed. All offloaded backups should be discoverable after object repository import procedure. If you believe that other product installation has written more objects to this repository, kindly bring this issue to our support team and let them confirm the behaviour experienced. Thanks!

[hoffmp]

Will do. I think I'm going to start from scratch. Build a new EC2 instance. Re-install Veeam, and attach the capacity tier now that I know the right way to do it. I made enough mistakes initially, that I'd like to have a clean replication of the behavior before I open a case. Thanks very much for your help with this

[veremin]

You are welcome. Please keep us updated on the results of your findings and feel free to open a ticket, if you need a further assistance. Thanks!

[hoffmp]

OK, I've think I've finally gotten to the bottom of this, and it's human error (mine). So here's what's going on. The SOBR did not have the "copy backups to object storage as soon as they're created" box checked. And to compound the problem, the Operational Restore Window was set to 0 days. Fortunately, the Backup Copy job that was writing to the SOBR has a very long retention period: 400 days. So all of the backups are present, they're just in the Performance Tier and never being written out to the Capacity Tier. So there's no data loss, which was my biggest concern. Next, what's the best way to fix this? If I enable the "copy backs to object storage as soon as they're created" and set the Operational Restore Window to something reasonable (like 1 week), will that offload the contents of the performance tier into the capacity tier? Or do you recommend another approach? Once again, thanks very much for your help and patience.

[veremin]

Next, what's the best way to fix this? If I enable the "copy backs to object storage as soon as they're created" and set the Operational Restore Window to something reasonable (like 1 week), will that offload the contents of the performance tier into the capacity tier?

Configuring operational restore window is required only for move policy, so if you want to copy existing backups (not to move only GFS restore points), you will need to enable "Copy backups to object storage" checkbox and that's is. Thanks!

[hoffmp]

Got it. Thanks! I'll do that now

[veremin]

Initial offload might take some time for 400 restore point long backup chain, but it eventually it should catch up. Thanks!

[hoffmp]

Oh yeah, it's still running ( https://photos.app.goo.gl/PAHQpcT1auM164w48 )

[veremin]

Sure, offloading 400 restore points will not be an immediate process, but as long as you have enough bandwidth it will be finished at some point. Thanks!

[johnomalley]

Correct. You don’t want to create a new sobr.

Just create the object repo and import the backups and start restoring away

This worked a Treat for me