How to Run Every Capacity Planning Scenario in Seconds, Part 2
In this post we will continue our examination of common scenarios you might need to plan for in your datacenter. If you are not already using Turbonomic, pull down the 30-day trial here and follow along! Click here to read Part 1.
With all the other projects and initiatives on the table, capacity planning can be just another monotonous task we don’t want to worry about. Let’s pick up where we left off and continue reviewing some more of our favorite capacity planning scenarios – and show how you don’t need to spend hours or days capacity planning. In this post we’ll focus on hardware configuration changes and business policies like DR and HA. As a quick recap, remember, the name of the game we’re playing is: What’s the optimal amount of hardware needed to assure application performance?
Scenario 4 & 5: Hardware Refresh or Hardware Upgrade
The Question: Should the hardware be upgraded (more RAM) or should it be refreshed? Or…
- …if the hardware is refreshed, how many new hosts should be purchased to maintain performance?
- …if the RAM is increased, does this assure performance to my current workloads?
- …if the RAM is increased, how many additional workloads can I safely support?
By upgrading your hardware or replacing them altogether, we struggle between the tradeoff of cost and performance. If we make the investment in new hardware, it’s likely we’ll need less of it to support our current workload demand (assuming this new hardware is more powerful). But how much less? I certainly don’t want to underspend and suffer performance issues. There’s also the cost tradeoff between new hardware and simply upgrading legacy hardware. Perhaps you can double up on RAM since that’s the constraining resource today.
Let’s head on back into the Plan tab. We’ve got legacy hardware that’s got to go – and we need to know how many new hosts to purchase. And maybe we’re debating between a few different hardware vendors with different specs. We can run a few different plans and compare. Let’s use the New Plan wizard to get started – click on the green plus button. Now choose, “Reconfigure Hardware.” Expand the Cluster drop down and select the cluster or clusters you would like to refresh. You can multi-select by using the Control and Shift keys. I’ll be upgrading my Lab cluster. Select “next” when done. You’ll notice the default selection on the left is “Replace Host Using a Template.” If you’re thinking about changing your Storage, you can select “Replace Storage Using a Template” as well. Click the “Physical Machines” group. Now we’ll use the Template dropdown to choose what hardware we’ll be upgrading to. Turbonomic comes with a few preconfigured hardware templates. You can select a model from the list, or scroll all the way down and select “Add New Template” to add your own. If you select “Add New Template” fill in the details of the hardware you would like to upgrade to. You can even input the cost of the hardware and have Turbonomic do some easy math for you. I’ll be using one of the preconfigured templates. Click “replace,” then “finish,” and now “run.” Let’s check out the results. With the particular hardware I’ve chosen, Turbonomic can keep the environment healthy on just 4 servers – which is 3 less servers than I have today. At the bottom I can see the total investment number based on the cost of each new host. You can now run this plan again choosing a different hardware configuration and compare.
Okay, so instead of refreshing the hardware, let’s run through a scenario where we upgrade hardware – like adding more RAM. The setup for this plan is very similar to replacing hardware. Let’s use the New Plan wizard to get started – click on the green plus button. Now choose, “Reconfigure Hardware.” Expand the Cluster drop down and select the cluster or clusters you would like to upgrade. I’ll be adding more RAM to the hosts in my lab cluster. Select “next” when done. Click the “Physical Machines” group. What we’ll be doing is selecting “Add new template” and building an identical server to what I have today, except for the RAM specs. If you don’t know the exact specs of your hardware today, you can use Turbonomic to find out. Navigate to the Inventory tab and find one of the hosts you’ll be upgrading. Expand the host and then expand the “Composed of” folder. By navigating to the Memory and PROC files I can see my hardware specs. For example, this host has 64GB of RAM and 8 Cores at 2.4GHz. Back to the Plan tab. My hosts’ RAM is going to be doubled to 128GB so I’ll plug in the appropriate value – notice that the units for RAM and CPU are in MB and Mhz respectively. Click “Create Template” and then select it from the drop down and “Replace”. Select Finish and Run. Using hosts with more RAM, Turbonomic will figure out how many are needed with that configuration. Let’s check out the results. If I upgrade to 128GB of RAM, Turbonomic can keep the environment healthy on just 4 upgraded servers. If I upgraded all 7 I would have even more capacity for growth. You can use the results and compare the cost to a complete hardware refresh, which is the plan we ran in the previous video.
Overall, we’ve reviewed five different, seemingly complex planning scenarios that can all be completed in a matter of seconds.
If you missed Part 1, give that a read here. Stay tuned for Part 3 where we’ll look at Failover and Disaster Recovery scenarios.