You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by ru...@albany.edu on 2003/01/01 20:10:13 UTC
RE: Remote JMeter
To Mike, Steve, Barry:
Hi Mike:
> As far as problems with remote testing, all I can say is that, on
> Windows2000, jdk1.4.0, it works for me. I won't deny that the classpath
> issues can be tricky, however.
Using JMeter 1.8? Would you mind sending your classpath and any related
environment variables you have set? Maybe this will help Steve as well?
Maybe I have conflict with other java paths already defined not
related to JMeter. Just don't know java well enough to debug this on my
own.
Hi Steve: Thanks for the sympathy. ;) Other team members here are in
dialog with Mercury Interactive if you can afford to go the
consulting route. They're _very_ expensive (at least when you
have no money). Estimate cost so far 50-70K:
http://www.mercuryinteractive.com/ but, I believe they're doing much more
than load testing.
Hi Barry:
> I have gotten JMeter working remotely. I run the GUI on a RedHat 8 box,
> andhave 5 Windows 2000 boxes running as slaves. Getting the CLASSPATH
> was a little tricky, and I put a "pause" in the .bat file...
Using JMeter 1.8 and jre 1.4.0 ? Please confirm since there seems
to be 3 big variances here, outside of classpath, which _could_ be
causing the problem (OS,JMeter, and JRE versions).
I did just try following the instructions first. Simply adding the 3
stated items to my classpath did not work. Did notice the missing .jar
extension on logkit. Had a pause in my script as well. ;)
Thanks for all the replies. Hoping to sort this out.
Happy New Year All!
-Ryan
---Start Whining [Read at own risk]:---
This is very frustrating since I've taken all the time to develop scripts
for testing yet I am now stuck with having a limited load I can apply. I
need to simulate 1000's of users on PC's which can probably only handle at
best 10-20 threads. Just not practical to scale up my load testing without
the remote testing portion working.
BTW: I had similar "java out of memory problems" as stated on the list
running 50 threads on a Win2k box with 512mg memory also encountered same
issue on PIII 450 128mg running RH 8 (20 threads if my memory serves). The
bulk testing machines I would have access to are only Win98 boxes with
64mg mem....
----End Whining.----
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: HTTP User Parameter Modifier
Posted by ms...@apache.org.
Combine the User Parameters with a StringFromFile or CVSRead
function and you can then stick your usernames and passwords into
a text file very easily.
-Mike
On 3 Dec 2003 at 13:37, russo@albany.edu wrote:
> Howdy,
> I see that the "HTTP User Parameter Modifier" has been deprecated?
>
> 1) Did it ever work?
> Using Win2k + Jmeter 1.9.1 + java version "1.3.1_01"
> where $JMHOME\bin\users.xml contains two parameters per
thread
> with paramnames "userid" and "password"
>
> I tried setting up a login request (recorded through the proxy)
which was
> modified to use the paramnames defined in users.xml ( ${userid}
and
> ${password} ). Pre-Processor "HTTP User Parameter Modifier"
was added to
> my Thread Group.
>
> The run fails to login.
> When viewing the request data (thru the Results Tree Listener) I
see the
> that ${userid} and ${password} have not been replaced. Copied
from the
> Results Tree Listener I get this:
> userid=${userid}&password=%24%7Bpassword%7D
> You can see the password value in my request "encoded" and
that the "${}"
> chars have been URL encoded.
>
> However, disabling the "HTTP User Parameter Modifier" and
adding "User
> Parameters" with appropriate values and re-running works...
>
> 2) What is the alternative given that I need to test several apps
which
> require logins with a load of several hundred users (and not
necessarily
> the same users). My test machines can only handle 50 threads so
I need to
> run my test plan on several machines; each with their own group
of user
> params.
>
> users.xml seemed like the perfect solution.... Having to manually
add 50
> sets of user name/value pairs to six separate copies of the same
test plan
> is awfully tedious and limiting when I can programatically add
user values
> to several users.xml files and distribute to each machine. It also
keeps
> these values separate from the test plan allowing me to have
several test
> plans using the same (or different users.xml).
>
> Ok, I've whined enough. Any suggestions?
>
> Regards,
> |----------------------------------|
> | Ryan Russo |
> | russo@[noSPAM].albany.edu |
> | University at Albany--Computing |
> | <<remove [noSPAM] to e-mail>> |
> | Technical Services Web Team |
> |__________________________________|
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-
unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-
help@jakarta.apache.org
>
--
Michael Stover
mstover1@apache.org
Yahoo IM: mstover_ya
ICQ: 152975688
AIM: mstover777
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
HTTP User Parameter Modifier
Posted by ru...@albany.edu.
Howdy,
I see that the "HTTP User Parameter Modifier" has been deprecated?
1) Did it ever work?
Using Win2k + Jmeter 1.9.1 + java version "1.3.1_01"
where $JMHOME\bin\users.xml contains two parameters per thread
with paramnames "userid" and "password"
I tried setting up a login request (recorded through the proxy) which was
modified to use the paramnames defined in users.xml ( ${userid} and
${password} ). Pre-Processor "HTTP User Parameter Modifier" was added to
my Thread Group.
The run fails to login.
When viewing the request data (thru the Results Tree Listener) I see the
that ${userid} and ${password} have not been replaced. Copied from the
Results Tree Listener I get this:
userid=${userid}&password=%24%7Bpassword%7D
You can see the password value in my request "encoded" and that the "${}"
chars have been URL encoded.
However, disabling the "HTTP User Parameter Modifier" and adding "User
Parameters" with appropriate values and re-running works...
2) What is the alternative given that I need to test several apps which
require logins with a load of several hundred users (and not necessarily
the same users). My test machines can only handle 50 threads so I need to
run my test plan on several machines; each with their own group of user
params.
users.xml seemed like the perfect solution.... Having to manually add 50
sets of user name/value pairs to six separate copies of the same test plan
is awfully tedious and limiting when I can programatically add user values
to several users.xml files and distribute to each machine. It also keeps
these values separate from the test plan allowing me to have several test
plans using the same (or different users.xml).
Ok, I've whined enough. Any suggestions?
Regards,
|----------------------------------|
| Ryan Russo |
| russo@[noSPAM].albany.edu |
| University at Albany--Computing |
| <<remove [noSPAM] to e-mail>> |
| Technical Services Web Team |
|__________________________________|
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Remote JMeter
Posted by ru...@albany.edu.
Possible solution:
I was able to get Remote JMeter to work after moving the JMeter out of
"Program Files" directory. This was discovered when my boss stepped in to
help and I ended up installing JMeter in c:\
I believe JMeter has a problem resolving its own classpaths when
directory path has spaces in it?
Proof:
I have two JMeter trees on the same box one in
C:\Program Files\jakarta-jmeter
and one in c:\jakarta-jmeter
Start cmd shell in C:\j2sdk1.4.0_03\bin
set JMHOME=C:\"Program Files"\jakarta-jmeter
set CLASSPATH=%JMHOME%\lib\ext\ApacheJMeter_core.jar;%JMHOME%\lib\jorphan.jar;%JMHOME%\lib\logkit-1.0.1.jar
start rmiregistry
In a new cmd shell
Starting C:\Program Files\jakarta-jmeter\bin\jmeter -s produces
the "..rmiregistry needs to be started error..."
Shutdown jmeter -s and exit cmd shell
In another new cmd shell start the version outside "Program Files"
Starting c:\jakarta-jmeter\bin\jmeter -s SUCCESS!
Note: I set the JMHOME every way I could think of without success.
That is, initially without quotes, then using the short name
progra~1, then quoting it. I'm guessing this is because rmiregistry is
picking these paths up fine and JMeter is attempting to construct its own
and failing when spaces are in the classpath.
For all those who had successes can you verify that you did not have
JMeter installed in "Program Files" or a directory path which contained
spaces? Or, point out a flaw in my theory...
Thanks For The FeedBack!
I doubt I would have stuck with this without it.
|------------------------------|
| Ryan Russo |
| University at Albany |
| Computing |
| Technical Services Web Team |
|------------------------------|
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: Remote JMeter
Posted by Joseph White <wh...@groton.cncdsl.com>.
I also had no problem running "remotely" when both instances were on the
same box (remote host = 127.0.0.1)
but it still wouldn't work across different machines. Running 1.8, JDK
1.4.1_01, Win2000. I could telnet
to the remote host.
<ru...@albany.edu> wrote in message
news:Pine.GSO.4.44.0301011312150.28944-100000@acunix1.albany.edu...
> To Mike, Steve, Barry:
>
> Hi Mike:
> > As far as problems with remote testing, all I can say is that, on
> > Windows2000, jdk1.4.0, it works for me. I won't deny that the classpath
> > issues can be tricky, however.
> Using JMeter 1.8? Would you mind sending your classpath and any related
> environment variables you have set? Maybe this will help Steve as well?
> Maybe I have conflict with other java paths already defined not
> related to JMeter. Just don't know java well enough to debug this on my
> own.
>
> Hi Steve: Thanks for the sympathy. ;) Other team members here are in
> dialog with Mercury Interactive if you can afford to go the
> consulting route. They're _very_ expensive (at least when you
> have no money). Estimate cost so far 50-70K:
> http://www.mercuryinteractive.com/ but, I believe they're doing much more
> than load testing.
>
> Hi Barry:
> > I have gotten JMeter working remotely. I run the GUI on a RedHat 8 box,
> > andhave 5 Windows 2000 boxes running as slaves. Getting the CLASSPATH
> > was a little tricky, and I put a "pause" in the .bat file...
> Using JMeter 1.8 and jre 1.4.0 ? Please confirm since there seems
> to be 3 big variances here, outside of classpath, which _could_ be
> causing the problem (OS,JMeter, and JRE versions).
> I did just try following the instructions first. Simply adding the 3
> stated items to my classpath did not work. Did notice the missing .jar
> extension on logkit. Had a pause in my script as well. ;)
>
> Thanks for all the replies. Hoping to sort this out.
> Happy New Year All!
> -Ryan
>
> ---Start Whining [Read at own risk]:---
> This is very frustrating since I've taken all the time to develop scripts
> for testing yet I am now stuck with having a limited load I can apply. I
> need to simulate 1000's of users on PC's which can probably only handle at
> best 10-20 threads. Just not practical to scale up my load testing without
> the remote testing portion working.
>
> BTW: I had similar "java out of memory problems" as stated on the list
> running 50 threads on a Win2k box with 512mg memory also encountered same
> issue on PIII 450 128mg running RH 8 (20 threads if my memory serves). The
> bulk testing machines I would have access to are only Win98 boxes with
> 64mg mem....
> ----End Whining.----
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: Remote JMeter
Posted by ru...@albany.edu.
Hi Scott:
> Not sure about the out of memory references that are included in this
>thread
Sorry about that. Should have sent a separate thread. As mentioned, I
don't feel all that comfortable with java yet.
> - I assume you are allocating more of the available memory to the JVM used
> to tun JMeter by altering the java mx parameter on startup?
> Scott
You assumed incorrectly. :) I incorrectly assumed java would use as much
available memory as possible. However, I guess that would break the
security sandbox? Anyway, thanks for the tip! If run into the problem
again I'll try:
java -Xms32m -Xmx128m given that I have 512mg.
Ref.:
http://bdn.borland.com/article/0,1410,28388,00.html#sec2-out-of-memory
-Ryan
> --
> Scott Eade
> Backstage Technologies Pty. Ltd.
> http://www.backstagetech.com.au
> .Mac Chat/AIM: seade at mac dot com
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: Remote JMeter
Posted by Scott Eade <se...@backstagetech.com.au>.
I have also had this working in the last couple of weeks - Win2k client and
Mac OS X server. One important step was to disable the firewall on the Mac
so that rmi could make it's connections. Other than this it worked fine for
me (a way of configuring the port that will be used is on my list of
potential enhancement requests).
Not sure about the out of memory references that are included in this thread
- I assume you are allocating more of the available memory to the JVM used
to tun JMeter by altering the java mx parameter on startup?
Scott
--
Scott Eade
Backstage Technologies Pty. Ltd.
http://www.backstagetech.com.au
.Mac Chat/AIM: seade at mac dot com
On 2/01/2003 6:24 AM, "Mike Stover" <ms...@apache.org> wrote:
> Yes, using 1.8. I don't know what all this talk of a pause in the script is
> about. I start rmiregistry on one command line, and then go start jmeter in a
> different one. I never use system-wide classpaths, as that's a recipe for
> hard-to-debug problems. instead, I define a more local classpath in the
> command line for rmiregistry, and then start rmiregistry. Admittedly, windows
> doesn't localize such things as well as L/Unix, but, that's how I do it since
> I'm more used to Unix in that regard.
>
> So, anyway, I just tried it again, and it worked perfectly my first try. For
> rmiregistry, I opened a windows command prompt, set the CLASSPATH to the full
> paths to jorphan.jar, logkitxxx.jar, and ApacheJMeter_core.jar. Then I
> started rmiregistry.
>
> Then, I opened a new command prompt, and changed directory to jmeter's bin
> dir, and started the server via "jmeter -s".
>
> then, I verified that my jmeter properties file included the "remote_hosts"
> line uncommented (it did, with ip of 127.0.0.1 - I have only one computer, but
> that shouldn't matter, it's still using RMI to communicate).
>
> Then, I started the JMeter gui, made a simple script,and ran it through the
> remote ip. No problems.
>
> I am running Windows2000, jdk1.4.0, and latest JMeter code from CVS, which is
> not different from 1.8 in this area.
>
> -Mike
>
> On 1 Jan 2003 at 14:10, russo@albany.edu wrote:
>
>> To Mike, Steve, Barry:
>>
>> Hi Mike:
>>> As far as problems with remote testing, all I can say is that, on
>>> Windows2000, jdk1.4.0, it works for me. I won't deny that the classpath
>>> issues can be tricky, however.
>> Using JMeter 1.8? Would you mind sending your classpath and any related
>> environment variables you have set? Maybe this will help Steve as well?
>> Maybe I have conflict with other java paths already defined not
>> related to JMeter. Just don't know java well enough to debug this on my
>> own.
>>
>> Hi Steve: Thanks for the sympathy. ;) Other team members here are in
>> dialog with Mercury Interactive if you can afford to go the
>> consulting route. They're _very_ expensive (at least when you
>> have no money). Estimate cost so far 50-70K:
>> http://www.mercuryinteractive.com/ but, I believe they're doing much more
>> than load testing.
>>
>> Hi Barry:
>>> I have gotten JMeter working remotely. I run the GUI on a RedHat 8 box,
>>> andhave 5 Windows 2000 boxes running as slaves. Getting the CLASSPATH
>>> was a little tricky, and I put a "pause" in the .bat file...
>> Using JMeter 1.8 and jre 1.4.0 ? Please confirm since there seems
>> to be 3 big variances here, outside of classpath, which _could_ be
>> causing the problem (OS,JMeter, and JRE versions).
>> I did just try following the instructions first. Simply adding the 3
>> stated items to my classpath did not work. Did notice the missing .jar
>> extension on logkit. Had a pause in my script as well. ;)
>>
>> Thanks for all the replies. Hoping to sort this out.
>> Happy New Year All!
>> -Ryan
>>
>> ---Start Whining [Read at own risk]:---
>> This is very frustrating since I've taken all the time to develop scripts
>> for testing yet I am now stuck with having a limited load I can apply. I
>> need to simulate 1000's of users on PC's which can probably only handle at
>> best 10-20 threads. Just not practical to scale up my load testing without
>> the remote testing portion working.
>>
>> BTW: I had similar "java out of memory problems" as stated on the list
>> running 50 threads on a Win2k box with 512mg memory also encountered same
>> issue on PIII 450 128mg running RH 8 (20 threads if my memory serves). The
>> bulk testing machines I would have access to are only Win98 boxes with
>> 64mg mem....
>> ----End Whining.----
>>
>>
>>
>>
>>
>>
>> --
>> To unsubscribe, e-mail: <ma...@jakarta.apache.org>
>> For additional commands, e-mail: <ma...@jakarta.apache.org>
>>
>
>
>
> --
> Michael Stover
> mstover1@apache.org
> Yahoo IM: mstover_ya
> ICQ: 152975688
> AIM: mstover777
>
> --
> To unsubscribe, e-mail: <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
>
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: Remote JMeter
Posted by ru...@albany.edu.
Hi Mike:
> Yes, using 1.8. I don't know what all this talk of a pause in the
> script is about.
I have everything starting from a batch file and wanted to give
rmiregistry time to load before calling jmeter -s. The plan was to create
a package on cd. The computers I have available are public user rooms and
I certainly wouldn't want to leave the software on the machines. Would
make a nice DoS attack... JMeter would be copied during testing to the PC
and Java would run off the CD. I even slapped a little java app together
to send the IP of the machine to the admin. who would be running the
JMeter client. That way I could just slap in a CD into each machine, run
the batch file and take off. After testing, run clean script and collect
CDs.
> I start rmiregistry on one command line, and then go
> start jmeter in a different one. I never use system-wide classpaths, as
> that's a recipe for hard-to-debug problems. instead, I define a more
> local classpath in the command line for rmiregistry, and then start
> rmiregistry.
I believe it would be a local classpath if run from batch file. No?
Anyway, I tried your method and still didn't work. :(
I included my steps below sig. if interested.
> Admittedly, windows doesn't localize such things as well
> as L/Unix, but, that's how I do it since I'm more used to Unix in that
> regard.
Agreed.
> So, anyway, I just tried it again, and it worked perfectly my first try.
Kick a guy when he's down huh? ;)
> I am running Windows2000, jdk1.4.0, and latest JMeter code from CVS, which is
> not different from 1.8 in this area.
Hmmm. Only difference is you're using the jdk and I'm using the jre. jdk
maybe come with libraries not in jre? I've downloaded jdk and will give
that a try.
Thanks For The Tips and Confirmation!
-Ryan
Unnsuccessful Steps taken::
Open cmd shell (1st attempt JMeter bin, then 2nd attempt in JRE bin):
Set a temp var. for ease of use
C:\j2re1.4.0_03\bin>set JMHOME=c:\Progra~1\JavaSoft\JMeter\jakarta-jmeter
Set the path
C:\j2re1.4.0_03\bin>set
CLASSPATH=%JMHOME%\lib\ext\ApacheJMeter_core.jar;%JMHOME
%\lib\jorphan.jar;%JMHOME%\lib\logkit-1.0.1.jar
Display the path
C:\j2re1.4.0_03\bin>echo %classpath%
c:\Progra~1\JavaSoft\JMeter\jakarta-jmeter\lib\ext\ApacheJMeter_core.jar;c:\Prog
ra~1\JavaSoft\JMeter\jakarta-jmeter\lib\jorphan.jar;c:\Progra~1\JavaSoft\JMeter\
jakarta-jmeter\lib\logkit-1.0.1.jar
(confirmed paths correct by copy pasting and performing directory
listing)
Start it
C:\j2re1.4.0_03\bin>start rmiregistry
Check RMI is running on default port
C:\j2re1.4.0_03\bin>netstat -a | find "1099"
TCP russo-dell:1099 russo-dell.acs.univ.albany.edu:0
LISTENING
Open cmd shell in JMeter bin dir. and run JMeter in server mode
(1st attempt opened fresh cmd shell, 2nd cd'd from the shell used to start
rmiregistry):
C:\Program Files\JavaSoft\JMeter\jakarta-jmeter\bin>jmeter -s
First line of log Reports:
01/02/2003 12:48:51 PM ERROR - jmeter.engine: rmiregistry needs to be
running to start JMeter in server mode java.rmi.ServerException:
RemoteException occurred in server thread; nested exception is:
java.rmi.UnmarshalException: error unmarshalling arguments; nested
exception is:
java.net.MalformedURLException: no protocol:
Files\JavaSoft\JMeter\jakarta-jmeter\lib\ant-1.5-optional.jar
However a check of RMI status again reveals:
H:\>netstat -a | find "1099"
TCP russo-dell:1099 russo-dell.acs.univ.albany.edu:0
LISTENING
TCP russo-dell:1304 russo-dell.acs.univ.albany.edu:1099
TIME_WAIT
jmeter still running. Start GUI, Open a script, attempt to run remote
script for 127.0.0.1:
JMeter GUI reports "Bad call to Host"
jmeter log now shows:
01/02/2003 1:05:25 PM ERROR - jmeter.gui: java.rmi.NotBoundException:
JMeterEngine
etc....
RMI again:
H:\>netstat -a | find "1099"
TCP russo-dell:1099 russo-dell.acs.univ.albany.edu:0 LISTENING
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: Remote JMeter
Posted by Mike Stover <ms...@apache.org>.
Yes, using 1.8. I don't know what all this talk of a pause in the script is about. I
start rmiregistry on one command line, and then go start jmeter in a different one.
I never use system-wide classpaths, as that's a recipe for hard-to-debug problems.
instead, I define a more local classpath in the command line for rmiregistry, and
then start rmiregistry. Admittedly, windows doesn't localize such things as well as
L/Unix, but, that's how I do it since I'm more used to Unix in that regard.
So, anyway, I just tried it again, and it worked perfectly my first try. For
rmiregistry, I opened a windows command prompt, set the CLASSPATH to the full
paths to jorphan.jar, logkitxxx.jar, and ApacheJMeter_core.jar. Then I started
rmiregistry.
Then, I opened a new command prompt, and changed directory to jmeter's bin dir,
and started the server via "jmeter -s".
then, I verified that my jmeter properties file included the "remote_hosts" line
uncommented (it did, with ip of 127.0.0.1 - I have only one computer, but that
shouldn't matter, it's still using RMI to communicate).
Then, I started the JMeter gui, made a simple script,and ran it through the remote
ip. No problems.
I am running Windows2000, jdk1.4.0, and latest JMeter code from CVS, which is
not different from 1.8 in this area.
-Mike
On 1 Jan 2003 at 14:10, russo@albany.edu wrote:
> To Mike, Steve, Barry:
>
> Hi Mike:
> > As far as problems with remote testing, all I can say is that, on
> > Windows2000, jdk1.4.0, it works for me. I won't deny that the classpath
> > issues can be tricky, however.
> Using JMeter 1.8? Would you mind sending your classpath and any related
> environment variables you have set? Maybe this will help Steve as well?
> Maybe I have conflict with other java paths already defined not
> related to JMeter. Just don't know java well enough to debug this on my
> own.
>
> Hi Steve: Thanks for the sympathy. ;) Other team members here are in
> dialog with Mercury Interactive if you can afford to go the
> consulting route. They're _very_ expensive (at least when you
> have no money). Estimate cost so far 50-70K:
> http://www.mercuryinteractive.com/ but, I believe they're doing much more
> than load testing.
>
> Hi Barry:
> > I have gotten JMeter working remotely. I run the GUI on a RedHat 8 box,
> > andhave 5 Windows 2000 boxes running as slaves. Getting the CLASSPATH
> > was a little tricky, and I put a "pause" in the .bat file...
> Using JMeter 1.8 and jre 1.4.0 ? Please confirm since there seems
> to be 3 big variances here, outside of classpath, which _could_ be
> causing the problem (OS,JMeter, and JRE versions).
> I did just try following the instructions first. Simply adding the 3
> stated items to my classpath did not work. Did notice the missing .jar
> extension on logkit. Had a pause in my script as well. ;)
>
> Thanks for all the replies. Hoping to sort this out.
> Happy New Year All!
> -Ryan
>
> ---Start Whining [Read at own risk]:---
> This is very frustrating since I've taken all the time to develop scripts
> for testing yet I am now stuck with having a limited load I can apply. I
> need to simulate 1000's of users on PC's which can probably only handle at
> best 10-20 threads. Just not practical to scale up my load testing without
> the remote testing portion working.
>
> BTW: I had similar "java out of memory problems" as stated on the list
> running 50 threads on a Win2k box with 512mg memory also encountered same
> issue on PIII 450 128mg running RH 8 (20 threads if my memory serves). The
> bulk testing machines I would have access to are only Win98 boxes with
> 64mg mem....
> ----End Whining.----
>
>
>
>
>
>
> --
> To unsubscribe, e-mail: <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
>
--
Michael Stover
mstover1@apache.org
Yahoo IM: mstover_ya
ICQ: 152975688
AIM: mstover777
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>