vRA/NSX deploy machines to compute cluster, edges to network cluster

[kallischlauch]

Hey,

i played a lot woth vRA and NSX in lab. but it was always the same cluster (compute and edges on same cluster)

now for a POC i have split the NSX in 2 clusters (edge and compute)

how can i make sure my vms land on the compute cluster and my edges (on demand nat, LB etc) on network cluster?

i guess the BG needs reservation in both

the machine needs to set its profile to a reservation matching the compute cluster, but will that not cause the edge to be deployed in compute as well?

any idea appreciated. i havent' tried it yet, just wondering

[z001yv0r]

Yes, you need to have separate Reservation Policy for Edge & Compute and the you need to select storage path instead of selecting the storage cluster in the resource tab of reservation.

Also select appropriate External-Profile like NAT, Routed.

[kallischlauch]

thx! i have seperate storage from compute and network cluster, so the VM will be easy. thx for that

how about the edge? how does vra/nsx know to deploy it to the network cluster?

[z001yv0r · Accepted solution]

As informed earlier create new reservation policy for compute and edge, then create separate reservation for compute and edge and map them into the appropriate reservation policy which you have created for compute abed edge.

Now create a blueprint and map the same edge reservation policy in network tab of routed gateway!...

[kallischlauch]

for those that are as slows as I am. in the blueprint properties: