Veeam backups to remote site

[mvindell]

Hello, we have purchased Veeam enterprise and are now finalizing our backup plans. I have some questions which I tried searching the forums for but did not find the answers too. Here is our configuration:

Primary site:
Vmware Esxi 4.1 (2 hosts) (20 VMs)
Iscsi SAN
Veeam Backup server on a physical box

Remote site:
We just purchased a new server for this which will have over 8 HDDs (2TB each).The remote site is connected to the primary site via a fiber connection (not to far away but enough for it to be safe). I once read that we could setup a dormant virtual environment on the remote site and using veeam, power on the backup files and go live in case of a disaster. Since we only have 1 server/host on the remote site how should we go about doing this? We had originally planned to shoot the veeam backups over to this remote site (no replication just backup). How could I go about setting up the desired plan?

my rational:
Setup the vmware hypervisor, then setup virtual disks and use those as the backup location for veeam, in the event of a disaster power on the backups. Wait we would need a backup veeam server then at the remote site?

Sorry if this has been asked before, I did not find any previous topic on this.

-Marco

[Vitaliy S.]

Hello Marco,

What will be your destination target for your backups? Virtual Disks?

Please be aware that it is not considered to be the best practice to store your backups on the virtual disks due to the reasons explained in this post.

Setup the vmware hypervisor, then setup virtual disks and use those as the backup location for veeam, in the event of a disaster power on the backups.

Could you please elaborate what do you mean by power backups on? Do you refer to Instant Recovery feature? How many VMs are you going to run from your backup storage?

Wait we would need a backup veeam server then at the remote site?

Not really. You can use Import backup functionality to import all the files to the new backup console and then perform a restore.

By the way why don't you want to use replicas instead?

Thanks.

[mvindell]

Yes I meant the instant recovery feature. I read the document you linked as well, thank you. The reasons we wanted to go with Backup instead of Replication is the use of compression and deduplication. These features are not available with replication correct?

Do you now of another way we could go about accomplishing this without using virtual disks (vmdks) as the target location for the backups? Using only one server I do not see how we can accomplish this.

We would use this primarily as a backup, incase of a disaster I wanted to use this as a backup virtual environment. We have about 20 VMs, in case of a disaster only about 5-8 would be needed to be powered on.

[Vitaliy S.]

These features are not available with replication correct?

These settings (deduplication/compression) are only applied to replica rollbacks (VRB files), because an original, unchanged .vmdk should be presented to a target host.

Do you now of another way we could go about accomplishing this without using virtual disks (vmdks) as the target location for the backups?

How about network shares?

On a side note, If you're going to run all 20 productions VMs simultaneously from the backup file, it might cause some latency and read/write issues, as it described right over here.

[mvindell]

I see, so with the exception of the first replication (the one that creates the .vmdk), all other replications will have compression/deduplication (the .VRBs). I had read this and thought otherwise:

Q: How exactly does replication work? How failover to earlier time is possible?
A: Replication is very similar to backup in reverse incremental backup mode, except that compression and deduplication are not used, and virtual disks are created in native format (as normal VMDK files). Failover to earlier time is enabled by storing earlier restore points (rollbacks) along with replica. For more information, refer to the following post, towards the end: Veeam Synthetic Backup Explained

Also, we plan to use separate jobs for a select number of servers. If we create separate jobs for the servers we plan on using it would create different backup files for each, in effect avoiding the issue of booting multiple VMs from a backup file.

Finally, I hope I am missing something huge, but what do you mean using network shares? If this is only one server, and vmware esxi hypervisor is running on it how would we go about creating network shares?

[Vitaliy S.]

Yes, all the rollbacks will be compressed and deduped, that is why in order to failover to a particular point in time you need to have a backup console presented for both replicas and backups.

Also, we plan to use separate jobs for a select number of servers. If we create separate jobs for the servers we plan on using it would create different backup files for each, in effect avoiding the issue of booting multiple VMs from a backup file

Ok. Give it a try to check out the performance on your DR site.

Finally, I hope I am missing something huge, but what do you mean using network shares? If this is only one server, and vmware esxi

Oh seems like I've missed that you're going to have only ESXi host on the remote site, nothing else.

[mvindell]

Thank you for your input vitality is it much appreciated.

[jroesner]

I'm curious to see how your implementation is going. We are doing almost the exact same here.
Duplicate servers, attached to a San, plus a physical server for veeam (all connected via 10G), then 100Meg fiber to a local server with oodles of diskspace that just handles replications. We're using replication to keep the 2 locations in sync with each other, plus then we take very critical systems, and move them to a remote office for piece of mind.

Are you considering moving data offsite (manually - tape, or removeable) as well?

[mvindell]

We are going to move weekly backups of our Oracle servers to a bank. Other than that, no we will not be moving anything to some sort of remote storage. We have decided to to use this other server strictly as a file server to hold all our Veaam backups and Zimbra backups, much like yourself.

[jroesner]

We're looking at replicating all the vm's to a backup vm server - that way they can be turned on "in case the meteor hits" as my boss puts it. (mind you the data centre is < 4km away so chances are it will take them out too..)

And now were testing to see what happens if we run a backup against the replicated servers, to move them to a removeable media.

[mvindell]

I wanted to go that route but the directors did not want to go that route. In order for that to work(backup VM environment, then just fire up the replications in case of a meteor lol) you would need at least 2 servers. That or a NAS/SAN and 1 server. The higher ups in my company decided having the backup VMs and critical data backed up properly would be enough...I explained recovery time and all that jazz. Good luck with your setup!

[jroesner]

We've got 2 Dell 715's at the DC connected to a SAN, plus the veeam server, then here we only have the 1 replicant server.

As most people have primary servers local, and replicant offsite, but due to environmental conditions we're the opposite, which is making it interesting for our offsite removeable backups.