What is the best way of protecting the Veeam Backup server?

[nestorius]

If my physical server corrupt. Can I recover my replicas via new server ?

[Andreas Neufert]

You mean Veeam VM replicas.
Yes, you can install B&R in a new server than import the config backup and from there you can replicate the VMs or failover again.

If you do not have the B&R Server at all, you can just start the VMs in the vcenter/esxi console. All restore points are kept as Snapshots. The only thing that you loose with this manual approach is that you can not failback the VM with Veeam then to original. But you can then treat the VM as a new VM and replicate it to any server including the original one.

[nestorius]

As I understand, I can start my replica vm as original vm, not failover-failback. Is this true ?

[DGrinev]

Any snapshot can be used to fire up a VM within vCenter or ESXi console, since the replica retains in a snapshot format, you can use it to register a VM as a new one (replacing the original if needed).
However, as Andreas wrote in the previous post, you will not be able to utilize failover-failback mechanism. Thanks!

[nestorius]

Thanks guys, all for yours useful informations...

[mkaec]

That's a very good design choice that the metadata is in the backup files. I recently found out that a trendy competitor doesn't do that and if you lose your backup server configuration, all the backup files are useless.

[iburnell]

I've inherited Veeam setup. We are located at Site A and most servers have been put into a Data Center - lets call it site B. Setup is Physical server running Veeam Backup and Replication console at Site A and Virtual Proxies at each site using attached MSA storage for disk backups and tape units at both sites with their own media pools

Have been asked question about what happens if our office (Site A with Physical Veeam Server) goes down i.e. resilience to move to a DR site and still use Veeam

Thanks

[PTide]

Ian,

Kindly check considerations in this thread. These two resources are also worth looking at:

- BP guide
- KB article

Thanks!

[cbraafhart]

Dear Community,

I am currently setting up a new backup environment using B&R 9.5 Update 4a and a question popped up in my head as followed:
If my current backup server, which handles all my current schedules, is offline are my backups being created?

And this immediately made me started to think on a redundant solution. What i would like to create is the following:

I have 2 Backup servers(VM) that have identical backup schedules for the sole purpose if one crashes or dies or gets corrupted the other one would kick in immediately without the problem of backups not being created due to the scheduler not sending the schedules to the designated servers.

Is this possible and how?

Thank you in advanced for any suggestions.

[lmbutault]

one solution could be to have proxy and repository roles on two physical boxes and run central server inside a VM.
That VM cloud be on external cluster or on local hyper-v (inside the box), clustered or réplicatged with veeam between two physical boxes.

regards

[foggy]

Hi Christopher, you can use configuration backup for that purpose. Please see above for details. Thanks!

[barrell]

Hi All

I've read a ton of post's and asked a couple of questions re Veeam B&R but sill can't fully understand the full solution. I have a 2 node ESXi cluster at Site-A that I wish to install Veeam B&R. This will be used to to copy data to Site-B over the WAN. I plan to run vMotion on the two node cluster and provide adequate resources to move VM's within this cluster to cover local redundancy.

If this cluster sustains a failure and the VM's are corrupt how do I restore from the remote site as my Veeam B&R will be installed as an appliance on Site-A cluster that is now down? Is this achieved using Veeam console on a local workstation at Site-A?

My original post for the minimum set of requirements is here vmware-vsphere-f24/minimum-componets-fo ... 59622.html

The model I now plan to use is as follows. One other question on the remote server, is this storage only or will I have to match the CPU/RAM of my ESXi cluster to cover the VM's being backed up? So in effect run ESXi at the remote site (not sure there is any benefit to this, possibly overthinking things?).


Many thanks.

[foggy]

Hi Lee, please review this thread for the possible ways of addressing your scenario. Thanks!

[bhagen]

We finally have our DR site running; six esxi 6.7u1 hosts, vsan, replicas of all our critical vm's being pushed to the dr cluster.

We have a single Veeam install on our production stack that's running all our backup, backup copy, replication, and quarterly tape jobs.

Obviously, if our production site goes down right now, we have no way to use Veeam to spin up the DR site. Its been suggested a couple times here and on the vmware forum that I now create a duplicate VBR install at the DR site. It's also been suggested that I run at least the replication jobs from that install.

Can somebody point me to documentation on how to accomplish this?

I mean, it's easy enough to clone my current VBR vm to the DR site, re-ip it and re-name it; but...is that the correct way to go about this?

Thanks!

[foggy]

The easiest way would be to set up a new Veeam B&R instance and import configuration backed up from the existing one.

[bhagen]

Thanks @foggy. So then I'll have duplicate vbr servers; would I just leave the one in the DR site running in case of an outage at the production site? I'm doing backup copies from prod to dr, then running replicas from the backup copies to the dr cluster. Should I be, as has been suggested, running the replication jobs from the vbr install at the DR site? If yes, how would I let the vbr install at the production site "know" that the DR install is running those jobs...just delete the jobs from the production install? But then, of course, the vbr installs wouldn't be identical.

Thoughts?

[foggy]

You can choose whichever approach is more suitable for you. Either have a single backup server in production (with periodic configuration backup to DR site to be able to quickly fire up a backup server there in case of losing the production one), or in DR site (always ready for immediate failover), or split the load between two instances and have the one in DR site responsible for replication jobs. In the latter case, please note that they will not be aware of each other and the load they assign to the shared proxy/repository servers. Another option is to use the DR instance to only replicate the main backup server instance and failover it as the first step of your DR procedure.

[bhagen]

Thanks @foggy! My coffee finally kicked in and I realized this topic has 10 pages...not just this page. So I'm reading thru it now, but your summary is excellent. I think for us the best option will be to "use the DR instance to only replicate the main backup server instance and failover it as the first step of your DR procedure".

Questions about that option:
1. In an emergency where the prod site is down, I'm guessing I'd log into the DR vcenter server client, and simply power up the replica of the prod VBR server? Then it's running on the new subnet, and I can use the failover plans (that I still need to create) to begin the failover of the prod vms?
2. Our prod VBR install is backing up veeam configs to a prod backup repository. I'm guessing I'd better back those configs up to the DR repository instead? So that if we spin up the DR site, we'll be able to import the most recent config, and also continue getting VBR config backups.

[skrause]

If you have a second B&R server in the DR datacenter, you setup the replication and failover plans targeting the main server in the production site.

Then you would just log into that server's console and initiate your failover plan.

If you have the backup jobs for the VM managing B&R in your production site on that server as well you likely won't have to worry about scheduling conflicts (though you might want to run the backup/replication jobs during a "non-backup window" where a lot of operations are going on, I backup my B&R servers at 10AM as an example.)

You can have the same repositories mapped to multiple B&R servers, they just won't automatically see all of the files created in them by jobs managed by other B&R servers.

[foggy]

I think for us the best option will be to "use the DR instance to only replicate the main backup server instance and failover it as the first step of your DR procedure".

Questions about that option:

1. If you decide to replicate the production Veeam B&R instance, then you just need to perform a failover of it (from the DR instance) and proceed with the plans.
2. If you back up configuration more frequently then replicate the server - yes. Otherwise, failover will be enough - replica will already have the latest configuration.

[bhagen]

Thanks again @foggy!

I've built a new vbr server on the DR stack. But when I try to add the DR backup repository, it fails, saying it's already connected to another vbr install.

Is there documentation somewhere that shows how to add the same repository to multiple vbr installs?

UPDATE:
Nevermind; seems my two vbr installs are different versions. I updated the prod install, and now I can add the same repository to the dr install.

[foggy]

If you're going to share the same server as repository between two Veeam B&R instances, just add it to both installations and create repositories pointing to different locations. Make sure both installations are running the same Veeam B&R version and keep this notice in mind.

[bhagen]

Yeah; I thought I was going to be clever and use the dr install to do *all* my replicas. My prod install does regular backups to the prod nas, then does backup copies to the dr nas, then does replication from the dr nas (the backup copy jobs) to the dr esxi hosts. I updated all components and both vbr installs to the latest (9.5.4.2866.Update4b), and added the DR nas to the dr vbr install. I created a replication job for the production vbr install vm (replicate from source vm) to the dr nas, and that works. Then I figured that since all the backup copy jobs are in the dr nas, I might as well run all those replication jobs from the dr vbr install.

Nope; the dr vbr install can't tell that I've run new backup copy jobs.

So, looks like my only option (unless I'm still missing something) is to replicate the production vbr install via the dr vbr install, and then failover to that in case of emergency. Not a problem; now I know!

[foggy]

If you want to replicate from backups (that are already copied to DR by the production Veeam B&R instance), you can still do this, though please note that it is just a workaround that is not officially supported.

[aungthetzaw]

If we have two veeam backup environment, can we backup veeam servers each other from both sites?

[wishr]

Hello Aungthetzaw,

Welcome to Veeam Community Forums and thanks for posting.

The recommended approach is to take configuration backups. I've merged your thread with the existing one.

Thanks

[Johnson]

Due to budget constraints, customer wants to install VBR 6.5 into a secondary domain controller.
Configuration backup is not applicable in this scenario as we need to backup the entire secondary domain controller.

In this case, is it possible for Veeam backup server to backup itself, which is also the secondary domain controller?

[DGrinev]

Hi Johnson,

You can recommend your customer to use Veeam B&R Community Edition, which supports all variety of modern features without any cost.

Thanks!

[foggy]

In this case, is it possible for Veeam backup server to backup itself, which is also the secondary domain controller?

Yes, this is possible, though such setup is not considered as best practice. Please review
this thread for more considerations.

[Tecbil]

What about using Veeam Agent to backup the B&R server when it is running as a VM?