Can you run the AWS Backup Appliance on-prem?

Hi,

These backup appliances were architected within their respective clouds such that you could deploy from the cloud marketplaces. Whilst I see no technical reason why it “couldn’t” run from on-prem, it has not been built to run in that way and would be unsupported.

The original architecture was that this could be managed independently in the cloud without a VBR or with a VBR via the plug-in, we’re seeing a shift towards this just being another VBR component so who knows what the future will bring…

Check the help page to see how it works but to me it needs to be in AWS to work most efficiently with best performance - https://helpcenter.veeam.com/docs/vbaws/guide/welcome.html?ver=70

@MicoolPaul at least for our use case, I think we’d like it more just being part of VBR.  We have a site-to-site VPN between on-prem and AWS, so we already treat AWS like vSphere except in someone else’s datacenter.

When initially setting this up, my hope was that I would give VBR access keys, and that would basically be the beginning and end of it.  I don’t really like that there seems to be an entirely separate interface just for Veeam for AWS.  But if I’m understanding you correctly, this is because Veeam for AWS used to be a completely separate thing?

Hi @Loosus456 thats correct that it was built as a standalone but with the VBR plugin to communicate to the AWS appliance now being Veeam’s recommended management approach we’ll see less development on the direct Veeam console within AWS as it’s just wasting QA effort doubling the testing.

As for your backup copy question, you have to specify from backups as the option for your backup copy job to find the AWS backups as per https://helpcenter.veeam.com/docs/backup/vsphere/backup_copy_vms.html?ver=120

@MicoolPaul sorry, I am still missing something.  When I try to create a backup copy job, I have only two options: “From jobs” and “From repositories.”  I don’t have a “From backups” option at all.

I think I am missing something fundamental here.

Check this link here and see the NOTE on the page - Step 3. Select Workloads to Process - User Guide for VMware vSphere (veeam.com)

NOTE

You can use From backups source only if you have an external repository added to the backup infrastructure and have at least one backup file stored in it.

@Chris.Childerhose I think I’m dumb or something.

So am I somehow backing up AWS workloads without a repository then?  Is my my current setup hosed or what?

Not hosed just probably not fully configured as required for backup copy jobs.  Check the links in this thread to go over everything and go from there.

I’m thinking you haven’t done the external repository step as you mentioned in your other topic you can only take snapshots which isn’t the case if you create some AWS S3 storage, add this to VBR as an external repository type, your primary backup job for AWS EC2 VMs can target the S3 storage and your backup copies can use the S3 storage as a source

I just went through the docs again, and I tried setting this up, and it seems to technically function, but this seems way too complicated and also kind of janky in terms of AWS security.  It also seems to require firing up EC2 instances, along with keeping a bunch of stuff in S3 buckets that we really don’t want, all of which is going to drive up our monthly costs.

I think we are just going to delete everything and revert back to agent-based backups.  I usually don’t have problems with Veeam, but this AWS solution seems overengineered for what little we are trying to accomplish with it.  In fact, I wouldn’t go this way at all, except every week we have random issues with our agent-based jobs locking up and freezing until we restart the VBR server. :\

Sounds like a plan.  Best of luck.

Hi @Loosus456 

The VBfAWS solution does certainly need more architecture, no denying that, but it is based around interacting with AWS in the best possible manner. For example, without backups on S3, you’ll need to ingest your servers into AWS on the fly during a recovery scenario. Have you tested how long it would take to restore all servers back into AWS based on your Agent-Based backups?

If you’ve got issues with the VBR server freezing it sounds like the best way forwards is to tackle this problem, have you got a support contract?

@MicoolPaul yes, we have active support, but after some back and forth with the support folks, we just kinda gave up and are trying to diagnose it on our own.  We just haven't been able to figure out the root cause yet.  It happens only about 15% to 20% of the time, but even that often requires we monitor backups pretty closely at least a few hours per week, which is a bit of a hassle.  We were hoping the AWS solution might be our fix, but it's just too much for us.

Regarding backups on S3: if we were to store our backups on AWS (which I think is a terrible idea by the way since that's where the production data is stored too), I am unsure why we would even need Veeam.  I'd just directly use AWS Backup.  It does the same thing, requires zero infrastructure, costs a lot less, and is built right into AWS.

Hi @Loosus456 - did the Agent backup solve your issue?  If so you can mark your post above as the accepted answer to this thread so others can see it.  If not let us know so if this is not something that can get answered we will adjust the topic to content instead of a question.

I'm not sure exactly what you're asking.  I told you that the Agent-based backups are failing?

Sorry I missed that part of the reply.  Not sure what else to suggest at this point other than a support case with Veeam to assist.

Like I said, we already did that but they gave up so we gave up.  As far as this thread: do whatever you want with it.  I don't care at this point.

@Madi.Cristil @safiya - can you convert this thread to content rather than question as there is no definitive answer here based on the last answer.