VCAP-DCA Mock Lab Q3
November 02, 2010 
Sean Crookston 
The following is a mock lab question for the VCAP-DCA exam. I’ve pulled some things from the blueprint that I felt were good to study and tried turning them into an individual question that would aid in studying for the exam. Please provide your feedback and answers to the question in the comments section below. I will post my solution and approve comments with the solution within a few days time.
Question
Your boss has come to you and finally admitted that he should have spent the money on VMware instead of using Hyper-V like you told him. As result he wants you to do the following:
- Setup a cluster of 3 ESX or ESXi hosts
- Ensure the cluster is highly available and that VM’s will recommended for placement on hosts intelligently in the event of a failure or during peak times where load on one of the hosts may be high.
- Set one of the three hosts to be used only for a failover.
- Your boss is also concerned with what may happen when one of the hosts becomes isolated from the other and is concerned with multiple instances of a virtual machine running at once.
Answer
Please provide your solutions, questions, etc in the comments section below.
Tags: 
Would build cluster with ESXi 4.1
Enable fully automated DRS
Enable HA with 1 failover host
Set isolation response to Shutdown
I like these — keep ‘em coming!
Damian definitely what I was thinking on doing with this one.
I believe that the right approach here would be:
Build 3 node cluster
Configure DRS to partially automated (the word recommended placement as listed in the question, makes it this way)
Configure HA with 1 failover host
Configure Isolation Response to Shutdown
at least that what I think
Rawlinson, Good point about the partially automated configuration based on the wording. I wrote the question and that even past me.
Although I can follow the reasoning behind Rawlinson’s post, I cannot see how ‘Partially Automated’ setting contributes to the cluster being highly available, as the scenario states.
After all, you might not be aware of the need to failover, just as you might not be aware of ‘peak load’ on one of the servers, in which case Partially automated won’t make you happy.
As far as I’m aware HA will focus on getting VMs up and running asap, normally followed by DRS ‘load-balancing’ (intelligent placement if you will) to relieve possible ‘peak load’ on individual servers. If needed DRS is even used before power-on VMs to ‘defragment’ the ‘failover-space’ available in the cluster in order to accomodate Admission Control.
As an aside, I hope the actual exam question wordings are not of the this type, where IMHO it really comes down to interpreting the english wording (semantics if you will), rather than dealing with the scenario-case.