Best Practice With New DR Site

[13Hemi]

I am looking at improving our backup infrastructure and adding a DR site. Currently, we have a single VBR server running on Server 2016 with a ReFS repo and monthly tape backups all in the same site. I've just received a new server and have a good opportunity to set things up a little better. The goal is not to have failover to the DR site but rather offsite backup and immutability. My thought was to set up both the prod and DR site with Linux repositories with the DR site having immutability enabled and running VBR/VDRO on a separate Windows host in the DR site rather than the prod site. Tape backups would continue to run in the prod site.

My question is - does what I'm thinking make sense? Would there be anything you would do differently? This is my first major DR overhaul so looking at best practice for everything.

Also, can a Linux repository also act as a tape server via SAS? Our current tape library is connected to the Server 2016 server via SAS. All tape backups are stored offsite from the prod DC.

[PetrM]

Hi, Jason and Welcome to Veeam R&D Forums!

Your idea to run VBR/VDRO on the DR site seems to be reasonable, according to our best practices it's recommended to keep a single Veeam B&R server on the DR site to provide seamless failover. Your production server may have a role of the tape server, thus the tape backups will continue to run on prod. Linux server can act as a tape server and SAS connections are supported. The OS requirements for the tape server are outlined on this page, the supported connection types are listed here.

Thanks!

[13Hemi]

Hi PetrM,

Appreciate your feedback! Is there a benefit of running a Linux Hardened Repository vs ReFS on Windows? Would you recommend having both prod and DR as LHR or just DR? I think I read that a server acting as a LHR is restricted to what roles it can run to just the proxy service. I don’t require immutability at the prod site.

[PetrM]

Hi Jason,

I strongly recommend LHR on the DR site to be sure that failover and restore operations are always available in case of any issue on the primary site including full production down and in case of malware activities or unplanned actions on the DR site.

Speaking about the Prod site, I'd suggest deploying LHR whenever possible but you could opt for Windows repository with REFS if you explicitly need block cloning for fast synthetics and space savings.

Thanks!

[JRRW]

@PetrM LHR doesn't preclude the use of block cloning - you can run XFS and LHR - best of both worlds, no?

[PetrM]

Hi Ryan,

That's a very good observation! I think that I should have mentioned it in my previous post. XFS is recommended for performance and space efficiency reasons, you can keep the backup chain in forward incremental mode with enabled synthetic full backups.

Thanks!

[13Hemi]

Quick question on this one with regards to this setup. With having the main VBR server located at the DR site per the recommendation, how would you recommend setting up a local mount server in the prod site?

According to the documentation at https://helpcenter.veeam.com/docs/backu ... deployment, it mentions:

For Linux, shared folder backup repositories and deduplicating storage appliances, Veeam Backup & Replication suggests the backup server.

Would I need a secondary backup server in the prod site? Is it recommended to be physical/isolated from the VM environment?

[PetrM]

Hi Jason,

I would recommend using a dedicated machine on the Prod site as a mount server for the Linux, shared folder repositories, and dedupe appliances. You can select an appropriate server in the repository settings. It might be whether virtual or physical machine, it's up to you to decide, there are no specific recommendations.

Thanks!