You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by Marcelo Jara <ma...@hotmail.com> on 2013/01/22 16:58:38 UTC
Performance penalty for using remote data files?
Hello,
I currently have several scripts that use large data sets (over 100K lines). The data files are currently stored locally on the Jmeter test server (Linux), but I would like to put them in a centralized location so that other teams can use them from their own machines. If I put them on a Windows share and then mount the share on the Jmeter server, would a test get poor performance due to fetching the next value from the network?
Re: Performance penalty for using remote data files?
Posted by Philippe Mouawad <ph...@gmail.com>.
You could use an OS Sampler:
http://jmeter.apache.org/usermanual/component_reference.html#OS_Process_Sampler
On Thu, Jan 24, 2013 at 9:35 PM, Philippe Mouawad <
philippe.mouawad@gmail.com> wrote:
> Agree with Hussey and Darren.
>
> Hussey Scott solution seems fine to me.
>
> https://twitter.com/philmdot <https://twitter.com/ubikingenierie>
>
>
> On Thu, Jan 24, 2013 at 5:21 PM, Darren Bown <da...@gmail.com> wrote:
>
>> Hi Marcelo,
>>
>> The short answer to this is "Yes, almost certainly". I've done rather a
>> lot
>> of heavy distributed load testing and any additional link in the chain is
>> likely to cause delays and therefore possibly distory results. Another
>> consideration is that depending on your network topology this could cause
>> congestion at other nodes and affect other users and indeed your own
>> tests.
>> I echo Scott, above, when he says that you should download your payload
>> and
>> execute with test data stored locally.
>>
>> Cheers,
>> Darren.
>>
>>
>>
>> On 24 January 2013 15:50, HUSSEY, SCOTT T <sh...@att.com> wrote:
>>
>> > In the past I've had a similar situation and used a step in a setup
>> thread
>> > group to download files to the local Jmeter engine.
>> >
>> > > -----Original Message-----
>> > > From: Marcelo Jara [mailto:marcelojara@hotmail.com]
>> > > Sent: Thursday, January 24, 2013 9:45 AM
>> > > To: user@jmeter.apache.org
>> > > Subject: RE: Performance penalty for using remote data files?
>> > >
>> > > Any insight from Jmeter developers would be greatly appreciated.
>> > >
>> > > > From: marcelojara@hotmail.com
>> > > > To: user@jmeter.apache.org
>> > > > Subject: Performance penalty for using remote data files?
>> > > > Date: Tue, 22 Jan 2013 10:58:38 -0500
>> > > >
>> > > > Hello,
>> > > > I currently have several scripts that use large data sets (over 100K
>> > lines).
>> > > The data files are currently stored locally on the Jmeter test server
>> > (Linux),
>> > > but I would like to put them in a centralized location so that other
>> > teams can
>> > > use them from their own machines. If I put them on a Windows share and
>> > > then mount the share on the Jmeter server, would a test get poor
>> > > performance due to fetching the next value from the network?
>> > > >
>> > > >
>> > >
>> >
>> > ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>>
>>
>> --
>> Darren Bown
>> tel:* +44(0)*7873 163 452
>> tel: *+44(0)7971 078 445*
>>
>
>
>
>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
>
--
Cordialement.
Philippe Mouawad.
Re: Performance penalty for using remote data files?
Posted by Philippe Mouawad <ph...@gmail.com>.
Agree with Hussey and Darren.
Hussey Scott solution seems fine to me.
https://twitter.com/philmdot <https://twitter.com/ubikingenierie>
On Thu, Jan 24, 2013 at 5:21 PM, Darren Bown <da...@gmail.com> wrote:
> Hi Marcelo,
>
> The short answer to this is "Yes, almost certainly". I've done rather a lot
> of heavy distributed load testing and any additional link in the chain is
> likely to cause delays and therefore possibly distory results. Another
> consideration is that depending on your network topology this could cause
> congestion at other nodes and affect other users and indeed your own tests.
> I echo Scott, above, when he says that you should download your payload and
> execute with test data stored locally.
>
> Cheers,
> Darren.
>
>
>
> On 24 January 2013 15:50, HUSSEY, SCOTT T <sh...@att.com> wrote:
>
> > In the past I've had a similar situation and used a step in a setup
> thread
> > group to download files to the local Jmeter engine.
> >
> > > -----Original Message-----
> > > From: Marcelo Jara [mailto:marcelojara@hotmail.com]
> > > Sent: Thursday, January 24, 2013 9:45 AM
> > > To: user@jmeter.apache.org
> > > Subject: RE: Performance penalty for using remote data files?
> > >
> > > Any insight from Jmeter developers would be greatly appreciated.
> > >
> > > > From: marcelojara@hotmail.com
> > > > To: user@jmeter.apache.org
> > > > Subject: Performance penalty for using remote data files?
> > > > Date: Tue, 22 Jan 2013 10:58:38 -0500
> > > >
> > > > Hello,
> > > > I currently have several scripts that use large data sets (over 100K
> > lines).
> > > The data files are currently stored locally on the Jmeter test server
> > (Linux),
> > > but I would like to put them in a centralized location so that other
> > teams can
> > > use them from their own machines. If I put them on a Windows share and
> > > then mount the share on the Jmeter server, would a test get poor
> > > performance due to fetching the next value from the network?
> > > >
> > > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
>
> --
> Darren Bown
> tel:* +44(0)*7873 163 452
> tel: *+44(0)7971 078 445*
>
--
Cordialement.
Philippe Mouawad.
RE: Performance penalty for using remote data files?
Posted by Marcelo Jara <ma...@hotmail.com>.
Thanks for the suggestion. Sounds like a good idea. Did you use a Java Sampler to do the copy?
> From: dazzlebown@gmail.com
> Date: Thu, 24 Jan 2013 16:21:43 +0000
> Subject: Re: Performance penalty for using remote data files?
> To: user@jmeter.apache.org
>
> Hi Marcelo,
>
> The short answer to this is "Yes, almost certainly". I've done rather a lot
> of heavy distributed load testing and any additional link in the chain is
> likely to cause delays and therefore possibly distory results. Another
> consideration is that depending on your network topology this could cause
> congestion at other nodes and affect other users and indeed your own tests.
> I echo Scott, above, when he says that you should download your payload and
> execute with test data stored locally.
>
> Cheers,
> Darren.
>
>
>
> On 24 January 2013 15:50, HUSSEY, SCOTT T <sh...@att.com> wrote:
>
> > In the past I've had a similar situation and used a step in a setup thread
> > group to download files to the local Jmeter engine.
> >
> > > -----Original Message-----
> > > From: Marcelo Jara [mailto:marcelojara@hotmail.com]
> > > Sent: Thursday, January 24, 2013 9:45 AM
> > > To: user@jmeter.apache.org
> > > Subject: RE: Performance penalty for using remote data files?
> > >
> > > Any insight from Jmeter developers would be greatly appreciated.
> > >
> > > > From: marcelojara@hotmail.com
> > > > To: user@jmeter.apache.org
> > > > Subject: Performance penalty for using remote data files?
> > > > Date: Tue, 22 Jan 2013 10:58:38 -0500
> > > >
> > > > Hello,
> > > > I currently have several scripts that use large data sets (over 100K
> > lines).
> > > The data files are currently stored locally on the Jmeter test server
> > (Linux),
> > > but I would like to put them in a centralized location so that other
> > teams can
> > > use them from their own machines. If I put them on a Windows share and
> > > then mount the share on the Jmeter server, would a test get poor
> > > performance due to fetching the next value from the network?
> > > >
> > > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
>
> --
> Darren Bown
> tel:* +44(0)*7873 163 452
> tel: *+44(0)7971 078 445*
Re: Performance penalty for using remote data files?
Posted by Darren Bown <da...@gmail.com>.
Hi Marcelo,
The short answer to this is "Yes, almost certainly". I've done rather a lot
of heavy distributed load testing and any additional link in the chain is
likely to cause delays and therefore possibly distory results. Another
consideration is that depending on your network topology this could cause
congestion at other nodes and affect other users and indeed your own tests.
I echo Scott, above, when he says that you should download your payload and
execute with test data stored locally.
Cheers,
Darren.
On 24 January 2013 15:50, HUSSEY, SCOTT T <sh...@att.com> wrote:
> In the past I've had a similar situation and used a step in a setup thread
> group to download files to the local Jmeter engine.
>
> > -----Original Message-----
> > From: Marcelo Jara [mailto:marcelojara@hotmail.com]
> > Sent: Thursday, January 24, 2013 9:45 AM
> > To: user@jmeter.apache.org
> > Subject: RE: Performance penalty for using remote data files?
> >
> > Any insight from Jmeter developers would be greatly appreciated.
> >
> > > From: marcelojara@hotmail.com
> > > To: user@jmeter.apache.org
> > > Subject: Performance penalty for using remote data files?
> > > Date: Tue, 22 Jan 2013 10:58:38 -0500
> > >
> > > Hello,
> > > I currently have several scripts that use large data sets (over 100K
> lines).
> > The data files are currently stored locally on the Jmeter test server
> (Linux),
> > but I would like to put them in a centralized location so that other
> teams can
> > use them from their own machines. If I put them on a Windows share and
> > then mount the share on the Jmeter server, would a test get poor
> > performance due to fetching the next value from the network?
> > >
> > >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Darren Bown
tel:* +44(0)*7873 163 452
tel: *+44(0)7971 078 445*
RE: Performance penalty for using remote data files?
Posted by "HUSSEY, SCOTT T" <sh...@att.com>.
In the past I've had a similar situation and used a step in a setup thread group to download files to the local Jmeter engine.
> -----Original Message-----
> From: Marcelo Jara [mailto:marcelojara@hotmail.com]
> Sent: Thursday, January 24, 2013 9:45 AM
> To: user@jmeter.apache.org
> Subject: RE: Performance penalty for using remote data files?
>
> Any insight from Jmeter developers would be greatly appreciated.
>
> > From: marcelojara@hotmail.com
> > To: user@jmeter.apache.org
> > Subject: Performance penalty for using remote data files?
> > Date: Tue, 22 Jan 2013 10:58:38 -0500
> >
> > Hello,
> > I currently have several scripts that use large data sets (over 100K lines).
> The data files are currently stored locally on the Jmeter test server (Linux),
> but I would like to put them in a centralized location so that other teams can
> use them from their own machines. If I put them on a Windows share and
> then mount the share on the Jmeter server, would a test get poor
> performance due to fetching the next value from the network?
> >
> >
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: Performance penalty for using remote data files?
Posted by Marcelo Jara <ma...@hotmail.com>.
Any insight from Jmeter developers would be greatly appreciated.
> From: marcelojara@hotmail.com
> To: user@jmeter.apache.org
> Subject: Performance penalty for using remote data files?
> Date: Tue, 22 Jan 2013 10:58:38 -0500
>
> Hello,
> I currently have several scripts that use large data sets (over 100K lines). The data files are currently stored locally on the Jmeter test server (Linux), but I would like to put them in a centralized location so that other teams can use them from their own machines. If I put them on a Windows share and then mount the share on the Jmeter server, would a test get poor performance due to fetching the next value from the network?
>
>