I suppose if you are testing an asynchronous system you don't have much choice. The only alternative is polling (I did something similar with a workflow app, but that does give you a way to retrieve the status of whatevers happening so that you can poll). regards deepak
On Mon, Nov 2, 2009 at 7:23 AM, Eric Smith <[email protected]> wrote: > I have done something like this in my testing. > In my case, the response is supposed to contain nothing but success or > failure. > But I wanted to know how long some stuff took (we do things *after* we > respond to the client.) > So in some ways I'm using jmeter to drive the service and do some > testing... and then use a perl script to validate even more via the log > file. > > I used scp to get the log files and run a perl script to validate the > output of the log files vs. my jmeter test logs (I wrote my own Simpler that > logs info.) It checks how long certain things take, correctness between > what we got back and what the log file said it should have been... and more. > > Since I was running purely in a QA environment - between a VM (jmeter) and > Solaris box (server) - I was allowed to set up passwordless ssh between the > systems and could copy the log files without any issues. I could also have > scripted supplying the password (not my top choice.) My other thought was > to share the log directory via NFS and just copy the log file via the mount > point. > > Eric > > -----Original Message----- > From: David Levine [mailto:[email protected]] > Sent: Sunday, November 01, 2009 8:07 PM > To: JMeter Users List > Subject: Is there a sampler that can read a tomcat log file? > > Hi, > > As part of my functional test, I'd like to assert that a particular line is > present (or not present) in the tomcat log file for my application. After > looking around at the available samplers and other JMeter tricks, it doesn't > look like there's a way to do this without writing code and extending > JMeter. Am I correct or is there a way to do this? > > OK, there's one kind of hacky way I thought of doing this, which is to > write a web service that returns the tomcat log as an XML document, and then > use an HTTP Request along with an XPath Extractor post processor, and look > for the log line I'm looking for. But that means I'd have to send back the > whole log file every time, which could be huge - so I don't like that. > > So I was thinking of writing a new sampler, called maybe the Tomcat Log > Sampler, that let's you specify a regex query for log lines that you're > looking for, and then returns those as XML, so that you can use the XPath > Extractor to pull the log lines you're looking for into a variable which you > can then subsequently use. The Tomcat Log Sampler would really just be a > proxy for a new web service I'd write, that would need to be running on the > tomcat web server. The Tomcat Log Web Service would perform the actual > query and return the results. That way it could scale. > > OK, does this make any sense or am I way off somewhere? > > David > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
