What-If in VMware vRealize Operations for VxRail/vSAN 6.7/7.0 Workload Placement

Posted by

I thought I would take a moment to write an article on an exercise that happened to me recently.  Recently, I was requested to see if a new project could go on an existing VxRail/vSAN cluster.  My trusty tool I like to go to for a quick answer, pretty visuals, and costs matrix is vROPs.  I ran the what-if for HCI and everyone was happy.  I did get the request to explain how vROPs came to the projected numbers.  However, my reply of it’s vROPs math was sufficient.  Below you will see the math formulas I used to reverse engineer vROPs Math.  Below my math formulas is the step by step how to in order to produce the report.

vRealize Operations Math:

As stated above.  I was requested to reverse engineer the math so my leadership could understand how vROPs came to these numbers.  I hope this helps.

vRealize Operations Memory Math

Cluster setting is annual growth of 12%  Current used space is 2.34TB, I am looking to add 754GB of memory at 70% utilization on Day 1 with a growth rate of 5%

2340×1.12=2620 —> projected utilization over next year prior to workload add

754x.05=37 —> growth

791x.7=553 —> 70% utilization

553+2620=3173 —> peak

I will be the first to admit I am my math is off by 43GB so my educated guess is that has to do with reporting current utilization vs what it may have been last time vROPs polled data.

vRealize Operations vSAN Math

Cluster setting is annual growth of 13%.  Current used space is 51.04TB, I am looking to add 806GB of storage at 70% utilized on Day 1 with a growth of 5%.  Plus, Dedupe/compression is enabled on this cluster at 1.3:1 ratio.

5104×1.13=5762  —-> projected utilized over next year prior to workload add
806x.05=40.3 —-> 5% growth over next year

846×2=1,692  —-> Raid-1 overhead

1692x.7=1184.4 —-> 70% utilized

1183/1.3=910 —> Post dedupe/compression

910+57620=58.53TB —> Peak utilization

As you will see below when we expand the view details section under Private Cloud: Datacenter the numbers will match up

Step by step workload report:

Machine generated alternative text:
vm vRealize Operations Manager 
Home 
Alerts 
Environment 
Administration 
Q 
BACK 
Quick Start 
Operations Overview 
Optimize Performance 
Workload Optimization 
Rightsizing 
Recommended Actions 
Optimize Capacity 
Overview 
Reclaim 
What-If Analysis 
Troubleshoot 
What-lf Analysis 
Overview 
Saved Scenarios 
Workload planning: Traditional 
Workload 
Workload planning: Hyperconverged 
c 
Workload 
Setup a scenario to 'deploy applications' by adding new workload or 'deprecate applications' by 
removing existing workload You can also evaluate the impact ot 'reclaiming inefficient workload' 
using remove workload scenario. 
by 
Plan tor 'future workload' to be deployed on vSAN environment with VMS associated with specific 
storage policy related factors (such as FTT, RAID), to evaluate if the usable capacity can cater to 
workload to be deployed. You can also evaluate the impact ot 'reclaiming inefficient workload' an 
vSAN cluster by using remove VMS scenario. 
ADD VMS 
REMOVE VMS 
ADD VMS 
REMOVE VMS

Home > Optimize Capacity > What-if Analysis > Workload Planning: Hyperconverged > Add VMs

Machine generated alternative text:
Workload Planning: Hyperconverged 
Add VMS to Hyperconverged Infrastructure 
SCENARIO NAME 
LOCATION 
Select datacenter

You’ll want to give it a name, I normally select the name of the project or the application that is making the resource request.

In location you need to select your vCenter you want to deploy the workload in.  Then select the cluster within the vCenter.

Machine generated alternative text:
APPLICATION PROFILE 
DATE 
RUN SCENARIO 
SAVE 
O Configure 
Select your workload 
CPU 
Memory 
Disk Space 
Expected utilization 
Annual Projected Growth @ 
Number of VMS 
v Additional vSAN configuration 
Account for Swap Space 
Host failures to tolerate 
Fault Tolerance Method 
Dedup 
Start date 
4/08/20 
CANCEL 
C) Import from existing VM 
v VCPU 
RAID-5 
End date* 
7/08/20 
ADVANCED CONFIGURATION 
ADVANCED CONFIGURATION 
•our engine is able to make projections 
up to maximum of one year from the 
current date

You have two options.
Option 1. Import from existing VM – this allows you to select VMs that exist in your environment assuming you will be deploying like for like VMs.

Option 2. Configure – This is the option I use majority of time because it allows me to get granular on net new workloads.  Once you fill in the values as granular as you desire you can click Run Scenario and get your results.

Machine generated alternative text:
Workload Planning: Hyperconverged 
Citrix Rebuild 
Scenario 
Add VMS to Hyperconverged Infrastructure 
Date 
Add VMS to: 
Total to be Added 13 V Ms 
Non-produc tion 
Memory 
Peak Disk Space 
58.48 TB 
PUBLIC CLOUD 
Microsoft Azure 
COST 
US$7,357/month 
LEARN MORE 
5278 ea 
EDIT 
Disk Space 
SAVE 
806 6B 
VIEW DETAILS 
x 
Apr 08, 2020 to Jul 08, 2020 
Richardson (DALPRDESXVICO' v 
CPU 
RUN SCENARIO 
163.42 GHZ 
Private Cloud: Datacenter 
The workload will fit. It will cost US$1,844/month*. Your time remaining is > 1 Year. 
• This cost calculation is based on monthly average utilization of workload for the selected 
The recommended cluster is: 
Non-production 
Public Cloud 
PUBLIC CLOUD 
Amazon Web Services 
aws 
COST 
US$7,713/month 
LEARN MORE 
peak CPU 
565.71 GHz 
PUBLIC CLOUD 
IBM Cloud 
COST 
US$5,335/month 
LEARN MORE 
Peak Memory 
3.13 TB 
PUBLIC CLOUD 
Google Cloud 
COST 
US$5,853/month 
LEARN MORE

If you click view details it will expand out into greater detail:
 

Machine generated alternative text:
O CPU* 
• Projected utilization: 398.01 GHz/770 GHz 
• With Added Workload: *163.42 GHz, 561.43 GHz/770 GHz 
•All capacity utilization numbers are projected peak values for forecasted perioc[ 
Memory* 
• Projected utilization: 261 TB/5.33 Ta 
• With Added Workload: *534.31 GB, 3.13 TB/5.33 TB 
Disk space* 
• Projected utilization: 57.62 TB/141.41 Ta 
• With Added Workload: *875.16 6B, 58.47 TB/141.41 TB 
CPU 
Memory 
Disk Space 
> 1 year remaining until Disk Space runs out 
125K 
100K 
75K 
SOK 
25K 
Mar 16 
Mar 23 
Mar 30 
Apr 6 
Apr 13 
Apr 20 
Apr 27 
May 4 
May 11 
May 18 
May 25 
Jun I 
Jun 8 
Jun IS 
Jun 22 
Jun 29 
Jul 6
Machine generated alternative text:
Private Cloud 
PRIVATE CLOUD 
Dallas 
COST 
US$1,741/month 
LEARN MORE

Summary:
As you can see if you don’t need to reverse engineer the math this process should go quickly.  vROPs is a powerful tool and I am a fan of all the what-if scenarios.  As always, I hope y’all find this helpful.

Leave a Reply