I'd really like to see things like CPU monitoring in JMeter as well.
Probably the most flexible way would be to enhance JMeter to get readings from remote
(or local) JMX (Java Management Extension?) beans.
There was a recent JavaWorld article with source code for getting CPU percentages
Have any of the maintainers considered using JMX or Phoenix for enhancing
manageability of JMeter?
Last fall JavaWorld had an article with some source code on getting CPU% statistics.
It would seem to be relatively easy to put it into a JMX bean and deploy it on the
server JVM (assuming it
I think I must have been bumped off this list because I haven't received any e-mails
since the 3rd of January.
**
This email and any files transmitted with it are confidential and
intended solely for the use of the individual
It's not a big deal, but I'm getting this warning in my log file:
12/26/2002 9:30:27 AM WARN - jmeter.util: ERROR! Resource string not found:
[running_test]
Apparently, it's missing from the org/apache/jmeter/resources/messages*.properties
files.
Jonathan
P.S. I have what I think is a much
Here is a much cleaner (IMHO) fix for the Assertion Results threading problem. Could
someone else try this out and see how it works for them? It works fine for me in both
GUI and non-GUI mode.
This fix also avoids the problem that Scott found with the previous fix when running
in non-GUI
--
Scott Eade
Backstage Technologies Pty. Ltd.
http://www.backstagetech.com.au
.Mac Chat/AIM: seade at mac dot com
On 19/12/2002 5:15 AM, Jonathan Carlson [EMAIL PROTECTED] wrote:
Success! The ORO Perl5Matcher is not thread-safe. I checked the mailing list
and found some comments about
Here are all the changes again, but with thread protection added for Perl5Compiler and
pre-compilation of all regex patterns. I included the full methods since it seemed
easier to do and because I added javadoc where there was none. Any news on getting
the nightly builds going again?
1) The
One small change that got missed before:
There is an area in the evaluateResponse method that looks like this:
if (!pass)
{
result.setFailure(true);
result.setFailureMessage(
Test Failed,
--
Scott Eade
Backstage Technologies Pty. Ltd.
http://www.backstagetech.com.au
.Mac Chat/AIM: seade at mac dot com
On 20/12/2002 4:14 AM, Jonathan Carlson [EMAIL PROTECTED] wrote:
Here are all the changes again, but with thread protection added for
Perl5Compiler and pre-compilation of all
Here's what I have so far...
I put a breakpoint into the ResponseAssertion#evaluateResponse(...) method. It only
breaks if the contains test does not pass.
I then used BeanShell to create a new matcher and a new compiler to test the response
string. The new matcher and pattern compiler find
Success! The ORO Perl5Matcher is not thread-safe. I checked the mailing list and
found some comments about Perl5Matcher needing to be accessed via a ThreadLocal
instance in mult-threaded situations.
This is what I did to ResponseAssertion to fix my problem:
1) I changed the matcher static
undesirable, IMHO.
Has anyone else run into this?
Jonathan Carlson
Minneapolis, Minnesota
P.S. My apologies for the silly disclaimer that my company keeps sticking to the
bottom of my e-mails.
**
This email and any files transmitted
Arghh, I had a NullPointerException in the code I wrote to cache the patterns.
So now we know what happens when an unexpected error occurs in the ResponseAssertion.
If you notice that again, it's probably some error in the jmeter.log that will tell
you about it.
[EMAIL PROTECTED] 12/18/02
OK, I fixed my problem. Michal sent me a similar issue he had. It was an array out
of bounds exception in the Perl5Matcher which is most likely due to the threading
issue with Perl5Matcher that will be fixed whenever my patch from earlier today goes
in.
BTW, it's looking like the
Done. If you have any suggestions for debugging this, I'd appreciate it. I can
recreate the problem at will and can rebuild from the source if printlns are needed.
Jonathan
[EMAIL PROTECTED] 12/17/02 04:48AM
Dnia pon 16. grudzien 2002 22:17, Jonathan Carlson napisal:
Thanks, I hope a file
I have found what may be an important clue, but it doesn't make sense to me. Because
I can not do finds in the ViewResultsTree I have to copy it into a Notepad file
(yes, I have to use Windows) to do searches on it.
When I tried saving it I got a message saying there was at least one Unicode
I tried getting the source code via CVS, but had problems possibly because of the
firewall here. I also tried downloading a nightly build and found that there hasn't
been one posted since Oct 27th. Actually, that wouldn't help me anyways since it
appears to be only a binary build.
Any help
Scratch that. I see that it comes with the distribution. Sorry!
-
I tried getting the source code via CVS, but had problems possibly because of the
firewall here. I also tried downloading a nightly build and found that there hasn't
been one posted since Oct
I've been noticing that in certain circumstances the Assertion Results doesn't find a
string that is there. After looking into it a little deeper it appears that the line
where the string occurs is very very long. Is it possible that it has a weakness in
this case?
Jonathan
Minneapolis,
One idea: Maybe it can be found on Sun's Java bug database. Maybe the just messed
something up in 1.4.1. They might have a workaround suggested.
It's worth a try.
Jonathan
[EMAIL PROTECTED] 12/11/02 08:16PM
I want to use some user defined variables, but when I attempt to add these
to the
. Is that correct?
Jonathan Carlson
Minneapolis, Minnesota
**
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received
that would do something like this or would I have
to write my own Config Element, or have I missed something?
Thanks in advance for any hints in this regard.
Jonathan Carlson
Minneapolis, Minnesota
**
This email and any files
want timed is in the first request and sometimes it's in the
subsequent request.
Thanks!
Jonathan Carlson
Minneapolis, Minnesota
**
This email and any files transmitted with it are confidential and
intended solely for the use
Hi Jordi, Thanks for your response. Having options isn't bad. But I, personally, am
more concerned about the appearance to the user than I am about the individual request
timings. I would suspect some others feel the same way. But, then again, if a
certain aggregate request was slow, it
the
results page, but I prefer keeping my plan tree as uncluttered as possible.
Thanks!
Jonathan Carlson
Minneapolis, Minnesota
**
This email and any files transmitted with it are confidential and
intended solely for the use
25 matches
Mail list logo
