DR site setup

[cdeal]

I know there are many post about this topic and I think I have read every single one. I have been using Veeam since 2013 and have always been disappointed that there is no single supported setup for a typical (and simple) Primary Site/Disaster Recovery Site setup. Please correct me if I'm wrong, but I have found nothing in the documentation that covers this. I have a setup now that has been working (with limitations) for the past 6-7 years, but the bottom line is, it's "unsupported" by Veeam and I would like to find a "supported" solution that works for us and what I think is a typical situation. I see MANY post on the forum that are the same as mine and have not seen a solution that will work for us and is supported. I'm hoping someone here can help.

My environment: Primary and DR site with VPN connectivity (1GB). Separate vCenter servers at each site. Production VM's at each site that require backup. HP Storage Snapshot capabilities at each site. Tape Library at Primary site. Currently have two separate physical Veeam Servers at each site with local storage for backups (step 4 below briefly describes how these are used to create replica's).

Requirements: Replication of production VM's from Primary to DR. Offsite copy's of backups to DR site. Tape Backup at the Primary site. Utilize Storage Snapshots, SureBackup and SureReplica for all jobs. Ability to quickly restore from backup when needed. Ability to fail-over to DR Replica's in case of extended outage at the Primary site.

Setups considered:
1. Veeam server at Primary site controlling all Backup, Copy and Replication jobs with proxy at DR site. (will not work for us)
2. Veeam server at the DR site controlling all Backup, Copy and Replication jobs with proxy at Primary site. (will not work for us)
3. Veeam server at Primary site controlling all Backup, Copy and Replication jobs with standby server at DR site used to restore Veeam config database when Primary site is offline. (this one has possibilities but also many unknowns).
4. Separate Veeam servers at each site with "Dummy" jobs at DR site to recognize copied backup jobs from Primary site in order to create Replica's from backup job files. (this HAS been working but is unsupported).

Scenario's 1, 2 and 3 will not work due to Surebackup/SureReplica jobs not working across sites and Storage snapshots at each site not working due to storage controllers being in isolated non-routed subnets (which is best practice). I think all these would work if I could select an off-site Veeam proxy server to handle Sureback/SureReplica jobs and Storage snapshot creation. From what I understand, both of these are handled by the Veeam Backup server only and cannot be done by a Veeam proxy. I haven't tested Restores. This could be an issue for any scenario (other then #4) that would require restores to go across the VPN.

Any suggestions are appreciated.

I have a case open as well (ID# 04205541).

[PetrM]

Hi Craig,

The recommended setup is to have a single VBR instance on DR site to orchestrate all jobs on both sites.
As far as I understand this kind of deployment does not suit you because storage system on Primary site can not be accessed from DR site.

I would recommend to setup a single VBR instance on DR site and to use a proxy on the Primary site to ensure data read from HP storage system.
In order to provide management traffic from VBR to the storage system on Primary site, you may try to configure routing for the corresponding subnets.
But in this case you partially ignore the best practice mentioned in the first post.
SureBackup job would work if you selected advanced configuration in the VirtualLab creation wizard.

Basically, the setup 4 might be an option as well.

Thanks!

[cdeal]

"I would recommend to setup a single VBR instance on DR site and to use a proxy on the Primary site to ensure data read from HP storage system."
This would work fine other than it would require routing setup for the storage system across the VPN, just so the VBR server at the DR site could orchestrate the storage snapshot. You would think this could be done by the onsite proxy at the primary site. It would have access since it's local to the storage. That would solve the problem. Also, Surebackup jobs would not work across sites for the same reason. Again, if the local proxy (at the Primary site) could be used for the Surebackup ping test, that problem would be solved as well. To me, both of these are a limitation of Veeam.

"Could you please clarify in more details the setup 4? How did you configure these "dummy" jobs"
1. Separate Physical VBR servers at each site, both have local storage for backups/Backup copies. VM Proxy servers at both sites as well for hot-add jobs.
2. Daily Backup and Backup copy jobs run on Primary VBR server. Backup Copy jobs have DR VBR server as target.
3. Dummy backup jobs are setup on DR VBR server mapped to backup copy files. Otherwise, DR server would not recognize the backups.
4. Replica's from backup jobs run daily with the Backup Copy job files as the source.

"why do you think that this setup is not supported?"
I think it's mentioned many times in the forums that it's officially unsupported to setup Dummy jobs to recognize backup jobs/files created by other VBR servers. It also was mentioned on a call with support (ID# 04178050) when I recently had an issue mapping Dummy jobs to the Backup Copies, due to recent code changes.

[PetrM]

Hi Craig,

If you map the "dummy" backup copy job to a backup produced by another backup copy job: it would not provoke any issues as long as this "dummy" job is disabled.
Mapping is performed to a backup residing on the repository and it does not matter where the job which produced this backup was created, it's transparent for the logic of mapping.

From my point of view, the setup 4 is fine but I'm wondering to know what kind of issues did you have?
The similar example is described on this help center page and I don't see a big difference between backup and backup copy in this scenario.

Regarding the setup 2: routing will be required to ensure management traffic from VBR on DR site to the storage system whereas data read can be performed by proxy located on the Primary site, please check the requirements here. SureBackup should also work fine across sites: if Veeam server has an access to the Primary site, it will be able to reach VirtualLab which must be deployed on the Primary site as well.

Thanks!