You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by Gr...@MDSUK.COM on 2006/03/03 12:37:38 UTC
parallel samplers and nested thread groups
Hi,
In my load I want to test the following scenario as JDBC requests.
As a one off the user logs in (there’s one JDBC request attributed to
this) they can then view either a level one, level two or level three
summary screen with (60% 30% or 10%) probability for each. Each summary
screen is built via an initial JDBC request/SQL statement (for the header)
followed by say 4 simultaneous JDBC requests.
I’ve got myself really confused trying to implement this as a JMeter
testplan.
I guess I can use a mixture of a once only controller and the throughout
controller to simulate the logon, and probabilistic ‘level ? summary’
.selection. However I’m thrown as to how I then implement the summary
screen samplers, i.e. the single request followed by 4 simultaneous
requests. At first I hoped to use a nested thread group, however from
previous threads I understand this isn’t possible.
Please tell me I’ve missed another nice trick.
Many thanks
Greg
Re: parallel samplers and nested thread groups
Posted by sebb <se...@gmail.com>.
The HTTP Sampler can generate subsamples when it downloads derived
resources - so that might give some ideas for how to return more than
one sample result.
This should also show the samples correctly in the Tree View Listener.
S.
On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> Thanks Sebb,
>
> I'm simulating hundreds of users (I guess I'll have to look into
> distributed testing) and this is just a small proportion of my test
> script; it will have additional thread groups with different use cases.
> I thought I may end up using other samplers (beanshell, Java request or
> Junit) like you said and wrap my four calls in the one sampler. I guess
> this means I would only get the one response time for the sampler with the
> four calls in. Any suggestions on how I could bring back the response
> times from the individual SQL calls and get JMeter to include these in its
> report, would be greatly appreciated.
>
> Many thanks
>
> Greg
>
>
>
>
> sebb <se...@gmail.com>
> 03/03/2006 17:49
> Please respond to
> "JMeter Users List" <jm...@jakarta.apache.org>
>
>
> To
> "JMeter Users List" <jm...@jakarta.apache.org>
> cc
>
> Subject
> Re: parallel samplers and nested thread groups
>
>
>
>
>
>
> I don't think that is going to be easy using JMeter
>
> If you are prepared to write some Java code, you could potentially use
> the BeanShell Sampler to create 4 threads to run the JDBC queries and
> wait for the result.
>
> Or indeed you could create a JUnit parallel testcase to do the work,
> and then test this using the JMeter Junit sampler.
>
> How many different users do you want to simulate? And how many
> requests will each make?
>
> S.
> On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> > Sorry if my mail was confusing.
> >
> > The test is simulating the JDBC calls being an EJB application, where we
> > have a logon page followed by one of three possible summary pages. The
> > logon page is built from one SQL statement, the summary pages 5
> > statements. For the summary page we have one initial SQL statement
> issued
> > on its own, then four simultaneous statements. I wanted to replicate the
> > JDBC behaviour in the test plan. I guess I won't be able to, because I
> > already have a thread and this can't issue simultaneous requests needed
> by
> > the 4 parallel statements used in the summary page. Is this right ?
> >
> > By the way thanks for the quick response.
> >
> > Greg
> >
> >
> >
> >
> >
> > sebb <se...@gmail.com>
> > 03/03/2006 14:18
> > Please respond to
> > "JMeter Users List" <jm...@jakarta.apache.org>
> >
> >
> > To
> > "JMeter Users List" <jm...@jakarta.apache.org>
> > cc
> >
> > Subject
> > Re: parallel samplers and nested thread groups
> >
> >
> >
> >
> >
> >
> > Not sure I understand the bit about 4 simultaneous JDBC requests - is
> > that for one login, or 4 separate logins?
> >
> > A given thread (user) can only do one request at a time.
> >
> > S.
> > On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> > > Hi,
> > >
> > > In my load I want to test the following scenario as JDBC requests.
> > >
> > > As a one off the user logs in (there's one JDBC request attributed to
> > > this) they can then view either a level one, level two or level three
> > > summary screen with (60% 30% or 10%) probability for each. Each
> summary
> > > screen is built via an initial JDBC request/SQL statement (for the
> > header)
> > > followed by say 4 simultaneous JDBC requests.
> > >
> > > I've got myself really confused trying to implement this as a JMeter
> > > testplan.
> > >
> > > I guess I can use a mixture of a once only controller and the
> > throughout
> > > controller to simulate the logon, and probabilistic 'level ? summary'
> > > .selection. However I'm thrown as to how I then implement the summary
> > > screen samplers, i.e. the single request followed by 4 simultaneous
> > > requests. At first I hoped to use a nested thread group, however from
> > > previous threads I understand this isn't possible.
> > >
> > > Please tell me I've missed another nice trick.
> > >
> > > Many thanks
> > >
> > > Greg
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
> >
> >
> > ________________________________________________________________________
> > This e-mail has been scanned for all viruses by Star. The
> > service is powered by MessageLabs. For more information on a proactive
> > anti-virus service working around the clock, around the globe, visit:
> > http://www.star.net.uk
> > ________________________________________________________________________
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: parallel samplers and nested thread groups
Posted by Gr...@MDSUK.COM.
Thanks Sebb,
I'm simulating hundreds of users (I guess I'll have to look into
distributed testing) and this is just a small proportion of my test
script; it will have additional thread groups with different use cases.
I thought I may end up using other samplers (beanshell, Java request or
Junit) like you said and wrap my four calls in the one sampler. I guess
this means I would only get the one response time for the sampler with the
four calls in. Any suggestions on how I could bring back the response
times from the individual SQL calls and get JMeter to include these in its
report, would be greatly appreciated.
Many thanks
Greg
sebb <se...@gmail.com>
03/03/2006 17:49
Please respond to
"JMeter Users List" <jm...@jakarta.apache.org>
To
"JMeter Users List" <jm...@jakarta.apache.org>
cc
Subject
Re: parallel samplers and nested thread groups
I don't think that is going to be easy using JMeter
If you are prepared to write some Java code, you could potentially use
the BeanShell Sampler to create 4 threads to run the JDBC queries and
wait for the result.
Or indeed you could create a JUnit parallel testcase to do the work,
and then test this using the JMeter Junit sampler.
How many different users do you want to simulate? And how many
requests will each make?
S.
On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> Sorry if my mail was confusing.
>
> The test is simulating the JDBC calls being an EJB application, where we
> have a logon page followed by one of three possible summary pages. The
> logon page is built from one SQL statement, the summary pages 5
> statements. For the summary page we have one initial SQL statement
issued
> on its own, then four simultaneous statements. I wanted to replicate the
> JDBC behaviour in the test plan. I guess I won't be able to, because I
> already have a thread and this can't issue simultaneous requests needed
by
> the 4 parallel statements used in the summary page. Is this right ?
>
> By the way thanks for the quick response.
>
> Greg
>
>
>
>
>
> sebb <se...@gmail.com>
> 03/03/2006 14:18
> Please respond to
> "JMeter Users List" <jm...@jakarta.apache.org>
>
>
> To
> "JMeter Users List" <jm...@jakarta.apache.org>
> cc
>
> Subject
> Re: parallel samplers and nested thread groups
>
>
>
>
>
>
> Not sure I understand the bit about 4 simultaneous JDBC requests - is
> that for one login, or 4 separate logins?
>
> A given thread (user) can only do one request at a time.
>
> S.
> On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> > Hi,
> >
> > In my load I want to test the following scenario as JDBC requests.
> >
> > As a one off the user logs in (there's one JDBC request attributed to
> > this) they can then view either a level one, level two or level three
> > summary screen with (60% 30% or 10%) probability for each. Each
summary
> > screen is built via an initial JDBC request/SQL statement (for the
> header)
> > followed by say 4 simultaneous JDBC requests.
> >
> > I've got myself really confused trying to implement this as a JMeter
> > testplan.
> >
> > I guess I can use a mixture of a once only controller and the
> throughout
> > controller to simulate the logon, and probabilistic 'level ? summary'
> > .selection. However I'm thrown as to how I then implement the summary
> > screen samplers, i.e. the single request followed by 4 simultaneous
> > requests. At first I hoped to use a nested thread group, however from
> > previous threads I understand this isn't possible.
> >
> > Please tell me I've missed another nice trick.
> >
> > Many thanks
> >
> > Greg
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________
Re: parallel samplers and nested thread groups
Posted by sebb <se...@gmail.com>.
I don't think that is going to be easy using JMeter
If you are prepared to write some Java code, you could potentially use
the BeanShell Sampler to create 4 threads to run the JDBC queries and
wait for the result.
Or indeed you could create a JUnit parallel testcase to do the work,
and then test this using the JMeter Junit sampler.
How many different users do you want to simulate? And how many
requests will each make?
S.
On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> Sorry if my mail was confusing.
>
> The test is simulating the JDBC calls being an EJB application, where we
> have a logon page followed by one of three possible summary pages. The
> logon page is built from one SQL statement, the summary pages 5
> statements. For the summary page we have one initial SQL statement issued
> on its own, then four simultaneous statements. I wanted to replicate the
> JDBC behaviour in the test plan. I guess I won't be able to, because I
> already have a thread and this can't issue simultaneous requests needed by
> the 4 parallel statements used in the summary page. Is this right ?
>
> By the way thanks for the quick response.
>
> Greg
>
>
>
>
>
> sebb <se...@gmail.com>
> 03/03/2006 14:18
> Please respond to
> "JMeter Users List" <jm...@jakarta.apache.org>
>
>
> To
> "JMeter Users List" <jm...@jakarta.apache.org>
> cc
>
> Subject
> Re: parallel samplers and nested thread groups
>
>
>
>
>
>
> Not sure I understand the bit about 4 simultaneous JDBC requests - is
> that for one login, or 4 separate logins?
>
> A given thread (user) can only do one request at a time.
>
> S.
> On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> > Hi,
> >
> > In my load I want to test the following scenario as JDBC requests.
> >
> > As a one off the user logs in (there's one JDBC request attributed to
> > this) they can then view either a level one, level two or level three
> > summary screen with (60% 30% or 10%) probability for each. Each summary
> > screen is built via an initial JDBC request/SQL statement (for the
> header)
> > followed by say 4 simultaneous JDBC requests.
> >
> > I've got myself really confused trying to implement this as a JMeter
> > testplan.
> >
> > I guess I can use a mixture of a once only controller and the
> throughout
> > controller to simulate the logon, and probabilistic 'level ? summary'
> > .selection. However I'm thrown as to how I then implement the summary
> > screen samplers, i.e. the single request followed by 4 simultaneous
> > requests. At first I hoped to use a nested thread group, however from
> > previous threads I understand this isn't possible.
> >
> > Please tell me I've missed another nice trick.
> >
> > Many thanks
> >
> > Greg
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: parallel samplers and nested thread groups
Posted by Gr...@MDSUK.COM.
Sorry if my mail was confusing.
The test is simulating the JDBC calls being an EJB application, where we
have a logon page followed by one of three possible summary pages. The
logon page is built from one SQL statement, the summary pages 5
statements. For the summary page we have one initial SQL statement issued
on its own, then four simultaneous statements. I wanted to replicate the
JDBC behaviour in the test plan. I guess I won't be able to, because I
already have a thread and this can't issue simultaneous requests needed by
the 4 parallel statements used in the summary page. Is this right ?
By the way thanks for the quick response.
Greg
sebb <se...@gmail.com>
03/03/2006 14:18
Please respond to
"JMeter Users List" <jm...@jakarta.apache.org>
To
"JMeter Users List" <jm...@jakarta.apache.org>
cc
Subject
Re: parallel samplers and nested thread groups
Not sure I understand the bit about 4 simultaneous JDBC requests - is
that for one login, or 4 separate logins?
A given thread (user) can only do one request at a time.
S.
On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> Hi,
>
> In my load I want to test the following scenario as JDBC requests.
>
> As a one off the user logs in (there's one JDBC request attributed to
> this) they can then view either a level one, level two or level three
> summary screen with (60% 30% or 10%) probability for each. Each summary
> screen is built via an initial JDBC request/SQL statement (for the
header)
> followed by say 4 simultaneous JDBC requests.
>
> I've got myself really confused trying to implement this as a JMeter
> testplan.
>
> I guess I can use a mixture of a once only controller and the
throughout
> controller to simulate the logon, and probabilistic 'level ? summary'
> .selection. However I'm thrown as to how I then implement the summary
> screen samplers, i.e. the single request followed by 4 simultaneous
> requests. At first I hoped to use a nested thread group, however from
> previous threads I understand this isn't possible.
>
> Please tell me I've missed another nice trick.
>
> Many thanks
>
> Greg
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________
Re: parallel samplers and nested thread groups
Posted by sebb <se...@gmail.com>.
Not sure I understand the bit about 4 simultaneous JDBC requests - is
that for one login, or 4 separate logins?
A given thread (user) can only do one request at a time.
S.
On 03/03/06, Greg.Peace@mdsuk.com <Gr...@mdsuk.com> wrote:
> Hi,
>
> In my load I want to test the following scenario as JDBC requests.
>
> As a one off the user logs in (there's one JDBC request attributed to
> this) they can then view either a level one, level two or level three
> summary screen with (60% 30% or 10%) probability for each. Each summary
> screen is built via an initial JDBC request/SQL statement (for the header)
> followed by say 4 simultaneous JDBC requests.
>
> I've got myself really confused trying to implement this as a JMeter
> testplan.
>
> I guess I can use a mixture of a once only controller and the throughout
> controller to simulate the logon, and probabilistic 'level ? summary'
> .selection. However I'm thrown as to how I then implement the summary
> screen samplers, i.e. the single request followed by 4 simultaneous
> requests. At first I hoped to use a nested thread group, however from
> previous threads I understand this isn't possible.
>
> Please tell me I've missed another nice trick.
>
> Many thanks
>
> Greg
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org