What is the best scenario to deploy Auto Deploy and requiere services?
I'll be using autodeploy for the first time in the lab, where I hope to mirror one of my production environments. In the past, I used the SAN to boot all the ESX hosts (28 host x 4 datacenters), but next phase of my deployment, I'll have more 2 times more ESX hosts and less time to deploy them all, so I think this time I need a bit more of automation to help me get through all the milestones without missing a deadline.
Even though I'll be building a lab very soon, I wont have much time to try various configurations. So here is a very high level description of how all the datacenters are configured. All the management servers, such as NTP, DNS, LADP, puppet. etc, are all running in RH, the only Windows servers in the environment are dedicated for vCenter and SQL. Dedicated Management cluster using Cisco UCS B200 blades. Multiple VLANS for inbound and outbound traffic.This is a stand-alone environment (None of that Windows AD).
So my questions is what are the best practices to set up the AutoDeploy services and dependencies.
The only thing that I have clear now is to set up AutoDeploy in its own Windows box, but what should I do with the DHCP and TFTP servers. Can I combine the two of them in one server? Can I make the AutoDeploy server also the DHCP and TFTP server without affecting performance.
What are the recommended vHardware resources needed for AutoDeploy to support 54 ESXi hosts
I try to keep all my designs simple for those admis who are going to be left behind to manage, patch and support the environments. I don't mind setting up two more servers if it will make a huge difference in performance, but if there is no real gain, what would be a good scenario that fits in this environment?