In addition to VMware Site Recovery Manager testing, I’ve also built a 2-site Zerto proof-of-concept environment in the sandbox, which actually spans two geographically separated sites for a real-world test, minus the production workloads.
I have just concluded SRM with vSphere Replication testing, and we’ve decided to trash the array-based replication option, due to cost, complexity, dependencies, etc, etc… and favor vSphere Replication for an apples-to-apples test agains Zerto Virtual Replication. Besides, after spending a couple of weeks with our SAN engineers troubleshooting Array-based replication to no avail (in addition to all considerations listed above), we felt it would be way too expensive to have the same storage in each site we’d like to protect or recover to and we’d also have to re-architect how we design datastores for vSphere. Hypervisor replication is much more agnostic, and furthermore, so is Zerto.
I will write up a comparison on SRM and Zerto as soon as my documentation catches up with me for anyone interested in a side-by-side comparison.
Anyhow, just as I did with SRM, here is a topology diagram of how I have deployed Zerto in the sandbox environment. How-to’s and overall comparison/issue tracking/notes will follow. Firewall ports between all entities are also included, however, I did not include system specs for any of the infrastructure seen below. I also did not include any of the Zerto Backup or Zerto Cloud Manager functionality, as this was not a requirement for the testing. I will include system requirements in a later entry.
You can download a copy of this diagram as a PDF.
Note: This can also be replicated if you have a single physical ESXi host to build virtual ESXi hosts and infrastructure beneath it. There are a number of resources on the web that will show you how to build a nested ESXi lab, this is just one of them, but it is detailed, and easy to follow along with, and has lots of pictures. In order to duplicate this though in a single host with nested ESXi and vCenter, you will need to be conscious of resource requirements.