VMware VCAP-DCA Prep – Deploying DRS Clusters

by Paul on November 4, 2010

This post will cover some topics found in Objective 3 in the VCAP-DCA exam blue print. One thing I would like to call out before getting into the post is there are a number of great sites out there that are generating a lot of awesome content pertaining to the VCAP exams. Here are a few that I’ve been using to help gather study materials:

Sean Crookston’s VCAP Index
Kendrick Coleman’s VCAP-DCA Landing Page
Ed Grigson’s VCAP-DCA Index

The following tasks and objectives will guide you through configuring a cluster, creating resource pools, and ensuring processor compatibility for features like VMotion.

Scenario

You have a vSphere environment containing three hosts; two of the hosts have the same Intel Xeon E5450 chip set and the third host has the new Intel Core i7 chip set. The two E5450 hosts have 32GB of RAM and the i7 host has 96GB of RAM. Each host has two CPU sockets. The E5450 processors are 3GHz each, and the i7 is 3.3GHz. Between the two E5450 hosts and the i7 host, there is a total of 74,400 MHz of CPU and 163,840 MB of RAM.

Task 1

Your manager is requesting that you configure a cluster of ESX hosts to make efficient use of resources and provide maximum system up-time. The VMs must be able to automatically migrate during times of extreme resource contention and must not impact guest OS uptime.

Objective 1:

Create a cluster and enable DRS so that any VM can be VMotioned from host to host without service interruption.

Hint:

After creating the cluster, enable Enhanced VMotion Capability (EVC) and select the appropriate settings to match the processor type of which you are adding. In this case, it’s the Core i7. At the automation level, select Fully Automated and adjust the slider to the second position (2). This will allow vCenter to migrate VMs to allow for significant performance improvements.

Estimated time to complete: 5 minutes, not including reboot.

Task 2

Your manager has informed you that your department will become responsible for guaranteeing service delivery backed by an SLA. The SLA will state that there are to be three separate tiers of service, each guaranteeing varying levels of CPU and memory. The three tiers are:

• Tier 1: Mission Critical / Production = Dedicate 40% available CPU and memory, allow for burst usage.
• Tier 2: Non-mission Critical / Production = Dedicate 25% available CPU and memory, do not allow for burst usage.
• Tier 3: Test/Dev = Dedicate 10% available CPU and memory, do not allow for burst usage.

Objective 2

Using the same configuration in Scenario 1 and the newly created cluster in Task 1, create three resource pools to reflect the three service tiers.

Estimated time to complete: 5 minutes.

Hint:

After completing Task 1, create the resource pools as outlined in the requirements. The resource pools total up to 75% of available compute resources. This will ensure that there is adequate capacity if needed for burst. This will also leave enough resources for the hosts.

Task 3

Your manager is concerned with power consumption in the datacenter and would like for you to enable power saving features. The systems must be able to automatically power themselves off when not in use and on as resources are required.

Objective 3

Configure the DRS cluster to use DPM to allow vCenter to automatically implement power management recommendations. Adjust the threshold to allow vCenter to power on hosts if resource utilization exceeds the target range and power off hosts if utilization becomes extremely lower than the target range.

Estimated time to complete: 5 minutes, not including BIOS configurations.

Hint:

First, verify that your vSphere hosts are compatible with the DPM feature. DPM uses Wake-on-LAN or WoL, iLO or IPMI to power on hosts. If your system doesn’t have iLO or IPMI capabilities, then the NIC that is being used for Service Console or Management Connection must be able to support WOL. Once verified and configured, edit the properties of your DRS cluster and set DPM to automatic and adjust the slider to the second position (2).

As always, feel free to leave comments with your notes or things you would do differently. For example, I’d like to hear if you would choose to use nested resource pools instead of individual resource pools.

Tagged as: certification, drs, vcap-dca, vcda410, virtualization, vmotion, vmware, vsphere