Azure Site Recovery - component limits

Question

Hi, we are looking at replicating approximately 500 VMs to Azure using the VMware / Physical classic architecture (we cannot access the VMware hypervisor directly so will deploy the mobility agent). We cannot run the Deployment Planner (for the same reason), and some of the functionality and descriptions of limits are causing confusion. Hopefully easy for someone to explain =).

1. core limits - I know we can request more but there's no mention of how much more. Has anyone with a default Enterprise limit of 350 asked for say 1500 cores across various sizes? Hoping that should be okay.
   2. Recovery Services Vault - the Azure Subscription Service Limits documentation in the Site Recovery section states 'Number of servers per Recovery Services vault' = 250. https://learn.microsoft.com/en-us/azure/azure-resource-manager/management/azure-subscription-service-limits#site-recovery-limits. However the ASR docs seem to indicate up to 1000/2000, without specifically stating on-premises VMs are included in that number.
   3. Configuration Server - the 'Set up recovery at scale' doc states 8vCPU, 16GB memory and 600GB cache disk will support up to 550 machines (based on not using this server as a process server as well). https://learn.microsoft.com/en-us/azure/site-recovery/vmware-physical-large-deployment#set-up-a-configuration-server. But the how-to guide says the same thing, then under sizing and capacity requirements states the same spec for <100 machines.

Am I correct in thinking the <100 is more about the Process Server? If it said that I'd be happy, but it only mentions the configuration server in this section. https://learn.microsoft.com/en-us/azure/site-recovery/physical-azure-set-up-source#sizing-and-capacity-requirements

We will have a 10GB ExpressRoute, and VMXNET3 NICs, so I'm thinking 1 x configuration server based on the 550 number (this would not be used as a process server), then based on the lower number above we could have up to 5 process servers to get better overall throughput. But if there is a 250 limit on the vault, we'll need 2 or 3 vaults as well.

Appreciate any feedback on this... thanks!

Answer 1

HI Simon,

I would follow the MS Docs as a starter point and start baselining against the best practices and recommended specs, having said that all the parameters and configuration are based on the other dependency components and for that discovery and inventory is important.

Configuration server spec is a standard starting recommended but varies according to the environment, I will suggest lot of planning and testing, including design decisions is required over here.

If you have access to the Microsoft Account Manager via your company I will suggest to get the MS TAM involved to discuss the requirements and probably someone from the MS should work with your requirements as this is a majore design review and decisions to be made on setting up appropriate environment.

Hope this helps.

==
Please "Accept the answer" if the information helped you. This will help us and others in the community as well.

Answer 2

Document date: 07/06/2022
https://learn.microsoft.com/en-us/azure/site-recovery/vmware-azure-configuration-server-requirements#sizing-and-capacity-requirements
16 vCPUs, 32 GB mem, 1 TB cache disk = 150 -200 machines

Document date: 09/11/2021
https://learn.microsoft.com/en-us/azure/site-recovery/vmware-physical-large-deployment#set-up-a-configuration-server
8 vCPUs, 16 GB mem, 600 GB cache disk = Up to 550 machines

That's a wild spec variation! When I'm relying on MS docs for a design, I'd rather they both said "Hey mate, we've got no idea how many machines this can handle... start here and scale as required" =)