Backup V Replication - placingof Veeam machine.

[deesloop]

OK.
I've used Veeam for a couple of years only to do Replication of VMs to a DR site.
Purely to have hot standby. And its worked no problem - the twice I've had to call on it.

I have a couple of issues.

Production network, 3 ESX hosts on a NEXSAN SAN.
DR site - 100Mb Connection to main site, 3 older hosts (servicable) on a Dell SAN.
DR site has the VM running Veeam - just an XP machine.

Should Veeam be running in Production and backing up to DR
Or running in DR and backing up to DR
DR is jsut sitting there doing nothing so having it in DR (as it is just now) reduces load on production
But is it slowing things down or does it make no difference at all - the netowkr is the bottleneck?

Also I'm wanting to get rid of Commvault which handles backup.
So many agents and nonsense. Absolutely no need as far as I am concerned.,

Where do I do backup to?
Do I move my XP vm to the production environment and add space to the SAN - £££

Or do I leave it where it is and run the backup against the Production VMS or the DR VMs and use the DELL SAN ££ or the cheap & cheerful Thecus NAS?

Also can I recover an entire VM from a backup.
It seems very small having run one - it's compressed I know, but it's 15% the size of the replica?
Just concerned something may go awry.

Plus I'd like to be have a rolling 5 year backup window.
Perhaps 3 months fully recoverable and then a couple of years worth of month ends, then a couple of year ends.
Not hugely sure how to set that up. Any help greatly aprreciated on any of the points.

[foggy]

Should Veeam be running in Production and backing up to DR
Or running in DR and backing up to DR

Darren, general recommendations are to have two Veeam B&R instances, the one in production responsible exclusively for backups and another one in DR to manage replication jobs. Thus, you will be able to perform operational restores right in your production site and in case of a serious disaster (primary site goes down, for instance), failover/failback operations can be smoothly performed from DR.

But these are just general recommendations and you can set up whatever you like/need, it mostly depends on your specific restoration requirements. So if you'll decide to go on with a single Veeam B&R instance, I'd recommend placing it in DR (for the reasons stated above) and set up proxy servers in the primary site (close to the primary storage) for the optimal performance of source VMs data retrieval.

Where do I do backup to?

Again, depends on where do you need to have them. Upcoming Veeam B&R v7 will allow to successfully implement what we consider best practices in terms of local and offsite backups archival. You can do backups locally and then smoothly transfer them offsite over the slow link even without putting additional load on production.

Also can I recover an entire VM from a backup.

Sure, full VM restore is one of the basic restore capabilities built-in Veeam B&R.

It seems very small having run one - it's compressed I know, but it's 15% the size of the replica?

Compression and deduplication of backup files is one of the basic advantages of Veeam B&R backups.

Plus I'd like to be have a rolling 5 year backup window.
Perhaps 3 months fully recoverable and then a couple of years worth of month ends, then a couple of year ends.

Veeam B&R v7's new Backup Copy job type will also have built-in GFS functionality, so you will be able to meet these requirements.