Recently I got chance to do a nested vCF 3.5/3.7 deployment in my lab and it was a great learning. Few friends of mine reached out to me to know more about VMware Cloud Foundation product as a whole and how we can get our hands dirty on it.
Through this series of articles, I want to share my experience with you on how to do a successful vCF 3.7 deployment in a nested environment.
What is VMware Cloud Foundation (vCF)?
As per VMware official documentation
VMware Cloud Foundation is an integrated software stack that bundles compute virtualization (VMware vSphere), storage virtualization (VMware vSAN), network virtualization (VMware NSX), and cloud management and monitoring (VMware vRealize Suite) into a single platform that can be deployed on premises as a private cloud or run as a service within a public cloud.
vCF helps you to deploy a true SDDC environment in your infrastructure by following the VMware Validated Design recommendations and makes the life cycle management of SDDC very easy. I will talk more about this in future article of this series.
Physical Hardware Specification
In my environment I have 3 physical supermicro (X8DT3) hosts. Each host have following specifications:
- 12 x 2.4 GHz CPU.
- 128 GB RAM.
- 2X10 GB NICs.
- 1 SSD (500 GB) and 2 HDD (2 TB)
So total resources in my physical cluster is 80 GHz cpu and 384 GB memory and 16+ TB of storage capacity. I am using vSAN in my physical environment.
My base infra is running on vSphere 6.7
- ESXi 6.7 EP 05 Build 10764712
- vCSA 6.7 U1b Build 11727113
With vCF, you can either go for a Consolidated Deployment Model where both Management and Workload domain can sit in same cluster or you can have them spread across 2 different clusters (Standard Architecture Model ).
For deploying management domain, vCF 3.7 requires 52 vCPU and 116 GB of memory, and a total of 6.9 TB of disk space, including 30% space reservation.
Note: You will find this calculation in vCF deployment parameter sheet (under Management workloads domain tab).
Physical Infra Changes
There are few changes that you need to do in your physical infra:
1: Enable fake SCSI reservations: Nested Esxi deployment fails if you are deploying it on top of vSAN storage in your physical infra. This can be easily fixed by running below one liner command on all physical esxi hosts contributing to vSAN.
esxcli system settings advanced set -o /VSAN/FakeSCSIReservations -i 1
2: Allow network traffic from multiple MAC addresses: If vSphere version in your base infra is at lower version than the vSphere version that will be deployed via vCF, then you need to set Forged Transmits, MAC address changes and Forged Transmits to Accept on the portgroups which will serve as uplinks for your nested Esxi.
Note: In vSphere 6.7 the MAC Learning functionality, eliminates the need of enabling promiscuous mode. This has been explained in this article by William Lam.
3: Enable DHCP for IP allocation to VXLAN network: You need to have a DHCP server (external or AD integrated) running in your environment. You need to create a scope for IP allocation to VXLAN network (as of now we can’t specify ip pool in vCF deployment config sheet). The number of IP’s in scope should be more than twice the number of hosts you have in your nested setup.
That’s it for this post. In next post of this series I will list down the DNS requirements for a vCF deployment.
I hope you enjoyed reading this post. Feel free to share this on social media if it is worth sharing