Vous êtes sur la page 1sur 6

Validation Test Plan: Running [Software Product] on

VMware Infrastructure
Goals
The following test plan outlines the specific steps and configurations to be tested in the
validation process for VMware Infrastructure. The goals are to quantify the likely
performance of [product name] and to confirm that support of such a configuration will
not introduce new risk to our mutual customers.
Outline additional goals/deliverables for this exercise below:
Application Configuration and Requirements:

This section provides details on the specific components of the application. !here
possible a graphical representation of the application should also be provided with
mappings of various functional requirements to specific hardware hosts or virtual
machines.
Application
Component
Software Version Hardware Requirements
Application Server Product Version, OS CPU, RAM, DISK
Database Server
Web Server
Virtual Macine Configurations
The test configuration consists of [n] VMware "#$ hosts each running VMware "#$
version [x.y] update [z].
%ote& 'onfiguration below is a sampling of hardware available at the VMware validation
lab list out additional configurations as per the requirements of the software product&
!ardware Configuration "#$%& Virtual Macine
DL 580
(.)*+,- ( quad.core Intel $eon
$/*01
Four virtual machines configured as
application servers, each with:
2(3+4 56M 2 v!"s
7+b "thernet card 2#$ memory
7 278+4 #6# drives % &thernet card
#6% %et6pp 96# *1(1 '(#$ storage

DL 380
*.28+,- ( quad.core Intel $eon
/*01
Four virtual machines configured as
)eb servers, each with:
*(+4 56M :v!";
(+b "thernet card *+,-.
7 278+4 #6# drives */etwor0.
#6% %et 6pp 96# *1(1 *1torage.


One virtual machines configured as a
data server with:
:v!";
*+,-.
*/etwor0.
*1torage.

460c
(.**+,- ( <ual core Intel $eon
0781
One virtual machine configured as an
agent with:
28+4 56M :v!";
(+b "thernet card *+,-.
( 278+4 #6# drives */etwor0.
#6% %et 6pp 96# *1(1 *1torage.

460c
(.** +,- ( <ual core Intel $eon
0781 One primary virtual machine with:
28+4 56M :v!";
(+b "thernet card *+,-.
( 278+4 #6# drives */etwor0.
#6% %et 6pp 96# *1(1 *1torage.


One client 2wor0load driver3 virtual
machine with:
:v!";
*+,-.
*/etwor0.
*1torage.

Total storage required =for example>
0 <isk ?artitions&
(0+4 for :O1;
211+4 for :application;
backups flat file storage page
file
201+4 for data files =actual
database si-e will be in the 01+4
range>
/0+4 for indexes
01+4 for logs
'encmar( )or(load
This section describes the workload used the load driver technology and the specific
functional transactions included.
The workload is based on :load testing tool version x.y;. The workload consists of the
following transaction scripts =provide relevant workload operations below>&
Script Name Frequency of Execution Percentage of Workload
reate /ew 4ic0et 5 per minute 2(6
,dd )or0info to 4ic0et 2 per minute '(6
7elete 4ic0et % per minute %(6
lose 4ic0et 8 per minute 2(6
&scalate tic0et 2 per minute %(6
hange "ser 9nfo 2 per minute :6
"pdate 1upport 4eam 2 per minute :6
<riving the load of the workload will be done using :x. virtual machines each with
:two virtual !"s and 2#$ of +,-.. The test usually runs for :2 hours, with '(
minutes of ramp up, '( minutes of ramp down, and 8( minutes of steady state execution
where data is to be collected;.
6dditional notes& :4he database will incrementally grow during these runs, but the
overall volume is consistent enough that performance is not expected to be effected by
the small growth of the database..
Test Cases
The following usecases will be tested corresponding to our :medium and large reference
configurations.. "ach configuration@workloads is described below& *for example each
configuration, ' user counts will be attempted, while resources and performance
characteristics are monitored and recorded.
<Medium> Configuration
4he medium configuration consists of several virtual machines: two )eb servers, two
application servers, a report server, and the database. 4he test cases to be run and the
data to be collected are as follows:
<User Count> Ave e!pon!e
"ime
CP# #tili$ation
<Database>
%emory
#tili$ation
<Database>
CP#
#tili$ation
<Apps tier
VM1>
2((
5((
;((
<Large> Configuration
4he large virtual machine configuration consists of ' )eb servers, ' application servers,
a report server, and the database. 4he test cases to be run and the data to be collected
are as follows:
<User Count> Ave e!pon!e
"ime
CP# #tili$ation
<Database>
%emory
#tili$ation
<Database>
CP#
#tili$ation
<Apps tier
VM1>
5((
;((
%(((
,dditional test cases 2if any3:
"*ptionall+& VMware Infrastructure Testing
In order to understand how [91< !roduct] works with higher level functionality of
VMware Infrastructure we will perform one or more of the following tests.
&%'are v%otion "e!ting
!hile running the workload VMware vMotion is used to eAecute manual migration of
the database virtual machine from one VMware "#$ host to another. <uring this test
response time and transaction rates are monitored and any observed slowdown in
performance is measured. 9ive such operations will be eAecuted and averages are then
determined across the five.
&%'are (i!tri)uted e!ource Sc*eduler +(S, "e!ting
<uring this test virtual machines are assigned to VMware "#$ hosts such that the
maBority of the load will be on one host. In the first test set the aggressiveness level of
VMware <5# to C'onservative.D #tart up the test and monitor how VMware <5# moves
virtual machines across a cluster to balance the load. Monitor transaction throughput as
well as '?E utili-ation of the various hosts in the cluster. Fou should see '?E utili-ation
balance across the hosts with little decrease in throughput. %ow run the same test after
setting the aggressiveness of VMware <5# to a substantially higher value. Typical
deliverables are '?E charts from VMware v'enter #erver that reflect the balancing of
the load during these tests.
&%'are -ig* Availa)ility +-A, "e!ting
5un the workload on a clustered resource pool and then do a hard shutdown of one host
of the configuration. %ote how the virtual machines come up on another host in the
cluster. 5estart the application components if necessary and functionally verify that the
application is working again. Measure the time it takes for the virtual machine to start
accepting work again. This work is often most interesting in the case of services that start
automatically at reboot such as !eb servers such that the application will automatically
be ready to work as soon as the virtual machine is restarted on another host.
)rap,up and -ocumentation
!here possible results of testing are documented in a standard format provided by
VMware. This will then be entered into a template for a deployment guide to be used as
a starting place for documenting the results of the testing. !here possible the virtual
machines used for testing should be archived for possible use in the future. 6ny
outstanding issues should be documented for further followup with either VMware or
other I#V development teams.
VMware. Inc/ 0123 !ill4iew A4e Palo Alto CA 51021 6$A Tel 788,179,5:80 ;a< 9=2,1:8,=223 www/4mware/com
Copyright 2009 VMware, Inc. !! right" re"er#e$. %hi" pro$&ct i" protecte$ 'y (.). an$ internationa! copyright an$ inte!!ect&a! property !aw". VMware pro$&ct" are
co#ere$ 'y one or *ore patent" !i"te$ at http+,,www.#*ware.co*,go,patent".
VMware i" a regi"tere$ tra$e*ar- or tra$e*ar- o. VMware, Inc. in the (nite$ )tate" an$,or other /&ri"$iction". !! other *ar-" an$ na*e" *entione$ herein *ay 'e
tra$e*ar-" o. their re"pecti#e co*panie". Ite* 0o+ 09123VM3%e"t34!an3454362