Prod/DR site - How many VBR servers & placement?

[haslund]

I saw Gostev's response in thread here http://forums.veeam.com/vmware-vsphere- ... 47-15.html

Gostev: We recommend managing replication jobs with a separate backup server located in the DR site (to ensure you can perform failover when the production site is down). You can simply disable parallel processing on that server?

To stop hi-hjacking of that thread I took Gostevs advice and setup a new thread.

Basically the situation I am working on is:

One production site.
One DR site.

Requirements are:
All VMs must be backed up to prod site every 24 hours.
All backups must be copied to DR site every 24 hours.
Business critical VMs must be continually replicated.
Important but not business critical VMs must be replicated every 24 hours to DR site.

Originally based on Gostev's advice above I planned to setup one VBR server at prod site handling backup jobs and backup copy job. Then deploy a second VBR server at DR site to handle replications. Business critical VMs as normal replication job and important vms as replication from backup copy job.
But then I stumbled on http://helpcenter.veeam.com/backup/80/v ... ackup.html which states:

Limitations for remote replica from backup
Backups that you plan to use as a data source must be created with a backup job configured on the same Veeam backup server where you configure the replication job.

The way I understand this I will be required to setup only one VBR server at the DR site, otherwise how will this work?

Thank you.

[Gostev]

Hi, Rasmus.

I will need to talk to the devs to confirm (will only be able to do this in 10 days), but now that I see this note, I vaguely remember that replication jobs may be looking up VMs in backup files using unique ID only certain backup server knows (instead of moRefs, which can be duplicate). Not 100% sure though, so will need to double check.

If that is confirmed to be correct, then to achieve your requirements, I see two options:
1. Two backup servers: in PROD site to manage backups/backup copy + in DR site to manage replications, replicate direct from PROD site (not from backups).
2. Single backup server in PROD site as a VM, replicate that VM to DR site with Veeam, failover by manually/script starting replica VM in case if entire PROD site goes down. Backup server manages everything, replication can be done from backups as planned.

On a side note, the recommendation you are quoting is from pre-v8 times, so I definitely did not consider any advanced v8-specific scenarios at the time (such as replication from backup files). So, don't worry about me stating this back then - new scenarios v8 opens may require new recommendations.

Thanks!

[SGalbincea]

2. Single backup server in PROD site as a VM, replicate that VM to DR site with Veeam, failover by manually/script starting replica VM in case if entire PROD site goes down. Backup server manages everything, replication can be done from backups as planned.

This is exactly how I am doing it with Prod/DR requirements almost exactly as in the OP at my primary client. The only difference is that I am using Zerto for replication of the VeeamB&R VM and other mission critical VMs (our RPO needs to be sub 15 seconds). I have physical proxy servers at both Prod (2) and DR (1) with tons of fast cores and RAM that actually handle moving the bits from SAN to DataDomain, then copy job to DataDomain at DR (DDBoost ROCKS) or SpecraLogic tapes for archiving. The beauty of the replicated B&R server is that in the event of a complete outage at HQ we are up and running at DR with Veeam and critical VMs in the time it takes the VMs to boot and are only ever dealing with a single Veeam server.