Topology advice - Veeam Server high availability

[Gav@GH]

I'm looking for topology advice on a new build I'm working on. Specifically, I'm looking for options for the installation of the physical Veeam server, with regards to it's high availability in the event of a disaster.

Considerations:
- The Veeam server is physical, will be running version 8 and has attached storage and a tape library
- Utilize the Veeam server for regular backups of the Tier 1 VMs to disk
- Utilize the Veeam server for End Of Month archiving to tape
- Ability to bring replica's online in the secondary site, in a disaster event.

The complication I'm weighing up, is the ability to have LOCAL backups (which are fast and avoid replication over the WAN) Vs the ability to have a Veeam server online in a DR event.

One obvious option is to simply have the Veeam server in the secondary site.
- On the plus side: Simplified management of Veeam replicas (and backups if required, although SRM copies would be suffucient)
- On the down side: Increased replication times and traffic over the WAN and lack of IT resources for hands on duties (tape management etc).

Is it possible to have the Veeam server in some kind of HA configuration?

For example:
- Having a secondary virtual instance, with connectivity to a central SQL store?
- Is it possible to use a proxy server role for disk storage and tape backups in primary site (allowing Veeam server to be in secondary site)?

Environment information (which is outlined in the attached picture http://i.imgur.com/g2UfdLr.jpg):
- All Tier 1 VMs with structured data (Exchange, SQL, Files) in the primary site. This is protected to a secondary site using vSphere SRM.
- Tier 2 VMs are also in the primary site (as well as some in a tertiary site). These are replicated to the secondary site using Veeam.
- Sites are approximately 1000km apart, with ~15ms latency.

[Gav@GH]

Throwing up another option here for consideration.

This scenario places the Veeam server in the secondary site, with local backups taken to separate LUN on SAN.

This has the benefit of fast local backups, longer term copies offsite and simplified management of Veeam replicas in a DR event (although I still have to resolve local IT hands for mounting tapes etc)

Can anyone confirm if the backup copies utilize change block tracking? (in the past when I've done this with local replicas, the offsite copies see them as a changed snapshot and don't use CBT)

[veremin]

The second approach seems like a best option, indeed. Once v8 is released, you will be able to utilize tape library in primary site, as it won't be required any longer to connect tape library directly to backup console.

By "backup copies" you meant backup copy job? The backup copy job is forever incremental; which means that after initial synchronization it sends only changed block to target location.

Thank.

[Gav@GH]

Thanks Eremin,

Just to flesh out the details further;

So with option 1 (Veeam server in primary site), are there currently no high availability scenarios for the Veeam server to fail over to the secondary site?

What would the installed components look like in option 2 - are you saying that with a (virtual?) Veeam and/or proxy server, the tape drive can be fibre attached directly to the SAN fabric?

[veremin]

So with option 1 (Veeam server in primary site), are there currently no high availability scenarios for the Veeam server to fail over to the secondary site?

You can create a small share disk in secondary site, attach it to backup console and use as a target for configuration backup. This way, if primary site goes down, you can re-deploy VB&R in secondary site, restore configuration backup, and start managing backup/replication jobs.

What would the installed components look like in option 2

It will be similar to proxy server. The devices connected to such servers will be visible via backup console.

Thanks.