Can I design a vSphere Metro Stretched Cluster(vMSC) to achieve high availability across data centers in an environment managed by vRealize?

[mikew123]

Hello,

Say I have implemented vRealize Suite as my provide cloud. To achieve high availability (for both vRealize components themselves and user VMs managed by vRealize ) across two metro data centers, can I implement vMSC in such an environment?

if yes, any guideline?

If not, how can I achieve high availability for environment managed by vRealize suite?

thanks a lot!

Mike

[daphnissov]

I assume by "vRealize Suite" you primarily mean vRealize Automation (although you posted in the vRealize Orchestrator forum), is that right? If you already have it deployed, what architecture are you running?

[mikew123]

I think Yes. I have just moved the post to vRealize Automation.

I have not deployed the software but in evaluation.

[daphnissov]

vRA does not support a geographically distributed deployment, however if you're doing vMSC it's likely your latencies are already fairly low to possibly have that work for you. You would need the fully distributed, enterprise deployment using load balancers and separate IaaS components.

[mikew123]

Thanks for your suggestion.

I was not able to find a guideline about how to install vRA across two data centers (yes, they are connected with low latency link). The only thing I can find is the description in this link (vRealize Automation Multi-Data Center Data Deployments ). BUt that is obviously not a high availability deployment across two data center.

Do I need install two completely separated vRA installations in two data centers? If that is the case, will I be able to manage the two data centers using one single web console? And will be a tenant's network across the two data centers?

Thanks

Mike

[daphnissov]

There isn't going to be a guide specifically for vRA + vMSC, so you'd have to treat it as a single data center installation and make provisions to constrain half of the components to one side. So it would be one single fully-distributed deployment (enterprise method) with load balancers.

[kdelgadovmw]

The Reference Architecture guide lays out minimum requirements for metro stretched clusters - < 5ms latency and > 1GB bandwidth between datacenters.

https://docs.vmware.com/en/vRealize-Automation/7.3/vrealize-automation-73-reference-architecture.pdf