Hi,
if you want to have 2 DC for redundancy, il means you shouldn’t use the same instance of VCF.
in VCF, admin cluster will works into a single site, meaning if you have problem, with site where admin cluster is, you will loose all management part. Not a good idea (vCenter server for workload domain will run into admin cluster).
So, thinking of VCF (at least until we support Brownfield ingestion), will not give you availability you are looking for.
for your case, 2 site = 2 vcf instance. Then for each workload domain you create « locally », you can decide to implement DR solution.
vcf will manage local infrastructure (1 vcf per site). And dr solution will manage VM which need to be replicated on other site.
so for you case, do a dr infrastructure exactly like you should have done it with an infra without vcf.
you need to think about network reachability and data/VM replication to put that in place.
with VCF 5.2, we will support « Brownfield ingestion » which will help in such case as you will be able to import a vCenter which is not working into you admin cluster. It will give you new case to support you infrastructure. Vcf (sddc manager) will then be able to manage both site without risk of loosing management boni case of a site failure.
so, if you want to start now :
- 2 vcf instance : 1 per site,
- Make you de plan just like you should have done it without vcf
- or wait for vcf 5.2 to use brownfield support to bike new architecture.
whatever, keep in mind that network bandwidth and latency is key to build a DR solution.