• EN
Solved

VDRO - vCenter Location or HA Configuration

dloseke
Userlevel 7
Badge +6

Hey all, getting ready to setup VDRO in the lab and production, but I have questions on the location of the vCenter server.  Currently, my vCenter is located in the primary site, but wondering if it should be instead setup at the recovery site, or if I should have HA or ELM configured so that there is a second vCenter setup at the recovery site.  ELM doesn’t sound as useful for this situation, but HA may work and there is low enough latency between the primary and recovery sites to support HA.  That said, I was hoping to the put the witness at our third site, but latency is higher to that location so that may not work.  Anyhow, I know I need to get my VBR backup server moved to the recovery site (was already on my list) but what’s considered the best practice here?

icon

Best answer by MicoolPaul 7 November 2022, 11:55

View original
Derek M. Loseke, Senior Systems Engineer | Veeam Legend 2022-2023 | VMSP/VMTSP | VCP6-DCV | VSP/VTSP | CCNA | https://technotesanddadjokes.com | @dloseke

6 comments

Chris.Childerhose
Userlevel 7
Badge +20

Not sure what is best practice, but this site might help as it has the architecture and deployment scenarios for Orchestrator - Architecture Overview - Veeam Disaster Recovery Orchestrator User Guide

I know that the Orchestrator server needs a connection to the VC in order to do the DR.

Chris Childerhose - VMCA2022 | VMCE2023 | Veeam Vanguard 7* | Veeam Legend 3* | vExpert 5* | VCAP-DCV/VCP-DCV | Object First Ace | Cisco Champion | Twitter: @cchilderhose | Blog Site – https://just-virtualization.tech
RK99
Userlevel 5
Badge

Great Initiative taken by @dloseke for spreading the awareness and clarifying all the doubts on the concept of “Veeam Disaster Recovery Orchestrator”

dloseke
Userlevel 7
Badge +6

Not sure what is best practice, but this site might help as it has the architecture and deployment scenarios for Orchestrator - Architecture Overview - Veeam Disaster Recovery Orchestrator User Guide

I know that the Orchestrator server needs a connection to the VC in order to do the DR.

 

Yeah, I read this but found it lacking...the actually architecture of the Veeam servers was decent, but how it interacts with the vCenter was somewhat lacking.  I was going to look up vMiss (Melissa Palmer) and I guess I still could, but it appears that she left Veeam earlier this year and I didn’t realize it.  Based on the presentations at the Summit, @AlecKing might be able to help?  I’ll do some more hunting and report back.  I should note that I’m talking about Orchestrating Failover, and not Restore or Storage Failover.  That said, for one scenario I have, I do have VM’s being copied to the recovery site, but the plan was to use the replicas.

Derek M. Loseke, Senior Systems Engineer | Veeam Legend 2022-2023 | VMSP/VMTSP | VCP6-DCV | VSP/VTSP | CCNA | https://technotesanddadjokes.com | @dloseke
dloseke
Userlevel 7
Badge +6

Okay, found the below post in the R&D Forums that Alec responded to…VDRO can have one vCenter or two.  But it didn’t answer if you have two vCenters if they can be in a HA configuration, nor if there is only one what the preferred location is for the vCenter.  I’m fairly certain that in the case of a single vCenter, having it at the recovery site would be helpful, but I’m not sure if it’s capable of running with the vCenter down in the event of the vCenter residing at the primary site….

https://forums.veeam.com/post419773.html

 

Derek M. Loseke, Senior Systems Engineer | Veeam Legend 2022-2023 | VMSP/VMTSP | VCP6-DCV | VSP/VTSP | CCNA | https://technotesanddadjokes.com | @dloseke
MicoolPaul
Userlevel 7
Badge +20

Morning!

 

vCenter HA is not the answer unfortunately, though it looks like it. This is because, and I quote:

“vCenter HA has been tested and certified to perform on physically adjacent ESXi hosts, and is not designed as a disaster recovery solution for locations connected over long distances. The replication traffic is too sensitive to disruption in these configurations and commonly becomes a point of failure, even after a successful deployment.  For these reasons, the VCHA network does not support being configured over a WAN topology." Source: https://kb.vmware.com/s/article/85579

 

You should ensure the vCenter responsible for recovery is outside of the fault domain you’re trying to protect as you require the survivability of DNS & vCenter for a successful VDRO experience 😊 So this could be a single vCenter instance that lives at the DR site, or a local vCenter instance for the DR resources that VDRO is aware of for recovery operations.

Michael Paul - #VeeamLegend | #VeeamVanguard | Veeam Certified Architect | VMware vExpert | https://micoolpaul.com | Twitter: @MicoolPaul | Mastodon: @micoolpaul@masto.nu | Bluesky: @micoolpaul.bsky.social
dloseke
Userlevel 7
Badge +6

Morning!

 

vCenter HA is not the answer unfortunately, though it looks like it. This is because, and I quote:

“vCenter HA has been tested and certified to perform on physically adjacent ESXi hosts, and is not designed as a disaster recovery solution for locations connected over long distances. The replication traffic is too sensitive to disruption in these configurations and commonly becomes a point of failure, even after a successful deployment.  For these reasons, the VCHA network does not support being configured over a WAN topology." Source: https://kb.vmware.com/s/article/85579

 

You should ensure the vCenter responsible for recovery is outside of the fault domain you’re trying to protect as you require the survivability of DNS & vCenter for a successful VDRO experience 😊 So this could be a single vCenter instance that lives at the DR site, or a local vCenter instance for the DR resources that VDRO is aware of for recovery operations.

 

Thanks Michael, I was beginning to wonder if HA was a great option here as I know it was originally intended for both to be at the same location.  So it sounds like the best bet is going to be either to have a single vCenter at the recovery site, two vCenter’s separate from each other (one for the primary site, one for the recovery site), or two vCenter’s (again, one at each site) but in ELM so that each vCenter can manage both/all sites in a single pane, correct? 

Sounds like to me, if you’re running something like an Essentials/Essentials Plus/Foundations license, the only option is to have the vCenter run at the recovery site, have two license kits and run them standalone - one at each site, or to have two Standard licenses to run two vCenters linked.

Derek M. Loseke, Senior Systems Engineer | Veeam Legend 2022-2023 | VMSP/VMTSP | VCP6-DCV | VSP/VTSP | CCNA | https://technotesanddadjokes.com | @dloseke

Comment


Terms & Conditions