Classic Design Questions

[Almero]

Good Morning Team

I am new to Veeam and need to design the classic small regional site .

Active Site >
3 ESXi Hosts with 6.52U2
1 vCenter Appliance 6.5U2
1 DELL 740XD full of SAS disks running VEEAM BnR 9.5U4
1 Mid Range DELL array for source VMs
30 VMs running

DR Site >
Active Site >
3 ESXi Hosts with 6.52U2 ( older kit )
1 vCenter Appliance 6.5U2
1 DELL DR4100 VEEAM Repository ( end of life , and not really part of the day to day business )
1 Mid Range DELL array for VMs and incoming replicas .
Only vCenter , and VEEAM running.
Small 10MB metro Ethernet for replication .

So looking at the deployment scenarios on VEEAM KBs listed here , what would be the best approach .
( https://helpcenter.veeam.com/archive/ba ... anced.html )

We don't have a contractual DR test requirement , so I am not looking at disaster avoidance ideas .
ALL I want is to do daily backups on source/primary/active site , and have the ability to have all VMs recover in the event we lose the entire active site in DR ( Classic approach )

Questions >
1> If I opt for Advanced setup .
One physical DELL 740XD VEEAM server at primary site with all the features installed ( PROXY , vNFS , DB , Repository , WAN accelerator )
Only a virtual VEEAM proxy on DR site .
VEEAM Replicas running from Primary to DR site , either from source VMDKs or Backup Repo ( not sure what is the best option )

If i lose the entire Primary site , with my physical VEEAM server , i cannot recover it in the DR site to do a controlled fail over.
Do I then need to visualize the some of the VEEAM components , so I can replicate the VEEAM server itself to DR , and recover it first in the event of DR ?
So just use the DELL 740XD as a repository in primary site .

Forums suggest you can recover VMs in DR without having the VEEAM management server there , by simply powering on replicas , but recommend yo don't do this , as you cannot fail back .

2> Opt for Distributed Deployment .
Primary Site has the DELL physical 740XD VEEAM server with all roles + separate VEEAM Enterprise Manager Windows 10 VM
Replicate the VEEAM Enterprize Manager Windows 10 PC using Veeam BnR .

DR Site has a Virtual VEEAM full blown VM ( with proxy and WAN accelerator ) .
In the Event I lose my entire primary site , can I just recover my Windows 10 Enterprise manager PC and recover the normal VMs using the DR/local/surviving VEEAM server ?
I guess not , documentation is not very clear on that .

Please share your experiences , what are the best options / considerations .
I want to be able to auto fail back , and don''t want to over complicate things where I need to use VCHA , or replicate major VEEAM components .
Also my question about replicating VMs from Source VMDK's , or should i opt to replicate from the already backed up VEEAM forward incremental schedule ?

Thank you in advance for any positive contributions .

[bdufour]

i would manage both sites from one vcenter and one veeam instance installed at the DR site, with at least one domain controller there. thats how im setup. in the event u lose the primary site u can fail everything over from the DR site. once the primary site comes up, u can failback your replicas. if the primary site is going to down longer than 1 week id consider doing a permanent failover, then when the primary site comes back up, replicate the changes back to that site using the production source vm as a seed.

i dont see much (if any) impact when replicating from source vms - to answer that question.

[Almero]

Thank you Sir , that makes perfect sense .

I GUESS , it would help if i just put one VEEAM PROXY at primary site , to streamline chatter for normal "local" backups

[bdufour]

you got it, have a proxy at the primary site to handle that traffic, veeam and vcenter are only going to send commands back to the primary site and the proxy will handle the rest. also, you want to place your metadata repo closer to the source for your replicas.

[Almero]

Thank you very much for the detailed assistance bdufour . Does not look like VEEAM forum allows me to allocate points for correct answer so please accept my sincere gratitude .

The last consideration i am playing with is our inter site pipe ( Primary to DR ) is not very big , so might want to place the single vCenter appliance in the primary site for enhanced user experience for day to day business.

To protect that VCSA a normal ESXi to ESXi replica would serve us well.
With same subnets usable on both sides , it recon this will work.
I want to stay away from VCHA as far as possible , as I don''t have a 3rd site for the witness.

[bdufour]

if you want to protect your vcsa, as in replicate it to a different site - make sure when you install it use the FQDN, not the IP address. if you install with the IP address you will not be able to change the ip address if you failover to a replica at a diff site under a diff network. see here - https://www.vcdx200.com/2018/04/what-is ... -pnid.html

OR like you were getting at, if the primary site is down - you can advertise/route the primary site networks out of the dr site (if you cant change the IP of the vcsa for some reason).

keep in mind, if you place the vcsa (vcenter) at the primary site, and the primary site goes down and all of your jobs are mapped through vcenter- you will not be able to failover to the replicas properly. thats why i suggest having veeam and vcenter at the dr site for the ability to failover/failback everything properly in the event that the main site goes down. i have a 20mbps mpls circuit btw our primary and dr site, and it works well. we once had 10mbps and that was fine as well.

[Almero]

Thank you sir , we will have 10meg line only but all VMs are Centos web servers with low rate of change so should be Ok .
Thank you for the guidelines regarding vCenter , I will opt to replicate ( fqdn based ) vcsa ESXi to ESXi from the VEEAM server at DR site .

[foggy]

Here's a good topic discussing some of your concerns.