OK, I see. For some such checks, you might be able to use the new BeanShell sampler - this can invoke arbitrary Java code, for example read one or more files, and return whatever it likes to JMeter. [BeanShell is in the CVS version of JMeter, not in 1.9.1]
Or one could run another instance of Tomcat (or even write a simple Java http server) that could perform such checks on demand, and use a JMeter HTTP sampler to retrieve the information when required. S. -----Original Message----- From: Karthik Viswanath [mailto:[EMAIL PROTECTED] Sent: 04 November 2003 05:09 To: JMeter Users List Subject: Re: Preserving 'logins' accross jmeter sessions The reason for using a shell script is that we need to verify whether the UI is actually doing what it is suppose to have done. For eg, lets say the UI screen involves restarting tomcat (or any application). As a verification, I need to check from the logs and find out if it actually restarted tomcat. What the httpd returns would be a response of 200 OK which would not validate the test. I will try the jmeter server approach and let you know the results. Thanks Karthik BAZLEY, Sebastian wrote: >Likewise, I'm sure it could be done, and I'm not saying it would not be >useful. > >But I question whether it is a good idea to run lots of JMX scripts in >separate JMeter invocations. >I assume that these scripts are quite short, and if so, there will be a lot >of overhead in starting the JVM and JMeter. That's what I meant by JMeter >not being suitable - sorry, I should have been more explicit. > >However, it's just occurred to me that it might be possible to use JMeter >server-client mode to solve that problem. Completely untested, but it might >work for low volume testing: > >- start a JMeter server instance >- use "jmeter -n -r -t testfile" from the shell script to send the script to >the server > >I'm not sure what context the JMeter server keeps at the end of s run, so >this might be a non-starter without some additional code. > >== > >I also should have clarified what I meant about the security of the login >process. >If the server accepts a login on the basis of a cookie which can be passed >from session to session, then it seems to me that there is scope for that >cookie to be stolen ... > >S. >-----Original Message----- >From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] >Sent: 03 November 2003 15:08 >To: JMeter Users List >Subject: RE: Preserving 'logins' accross jmeter sessions > > >But how hard would it be? Why couldn't the cookie manager >serialize itself to a file and be reloaded by a different cookie >manager? If the manager is given a file, then they would all load >state at the start of the test from the file, and at the end, dump their >state. Synchronize file access and treat the file contents as a >queue/stack of cookie state (when a manager reads state, it also >deletes that entry). This could allow state to be carried over >between jmeter runs and between threadgroups that are running >serially. > >Since this isn't the first time it's been asked for, it seems worthwhile >to do. > >-Mike > >On 3 Nov 2003 at 12:46, BAZLEY, Sebastian wrote: > > > >>IMO, JMeter is not really designed for such usage ... >> >>Can you not use Response Assertions to do the checking instead >> >> >of using the > > >>shell? >> >>Could you not build up a JMX file to contain all the tests? >>It is possible to use functions and variables to parameterise tests >> >> >(see my > > >>posting in the thread RE: Using csv data as input) >> >>== >> >>As to avoiding logging in each time, that depends entirely on the >>application you are testing, and how it determines whether you >> >> >are still > > >>logged in. >> >>If it is not very secure, you might be able to extract one or more >> >> >tokens > > >>(e.g. cookies or URLs) from the initial login session, and pass >> >> >them to the > > >>next test. >> >>S. >>-----Original Message----- >>From: Karthik Viswanath [mailto:[EMAIL PROTECTED] >>Sent: 03 November 2003 07:34 >>To: [EMAIL PROTECTED] >>Subject: Preserving 'logins' accross jmeter sessions >> >> >>Hi, >> >>I am trying to develop a test suite which involves execution of a >> >> >'jmx' > > >>script followed by various verification steps for each testcase (in >> >> >a > > >>shell script) Currently I am using the command line options to >> >> >execute > > >>the jmx script. >>However, when the jmx files are executed in a batch, it takes a >> >> >long > > >>time to execute since >>1. I am invoking jmeter from the command line for every jmx file >> >> >being > > >>executed >>2. I need to login for every testcase execution. (Our UI uses >> >> >sockets to > > >>detect active users) >>A sample batch would look like >> >>./jmeter -n -t 1.jmx -l log.jtl -H <proxy> -P 1080 >>./1.sh >>./jmeter -n -t 2.jmx -l log.jtl -H <proxy> -P 1080 >>./2.sh >>... >> >>I would also like to know of any way to execute 2.jmx "without >> >> >logging > > >>in" again. >> >>Also any suggestions for a better approach than what I am >> >> >currently > > >>using would be very helpful. >> >>Thanks >>Karthik >> >> >> >> >> >> >>--------------------------------------------------------------------- >>To unsubscribe, e-mail: jmeter-user- >> >> >[EMAIL PROTECTED] > > >>For additional commands, e-mail: jmeter-user- >> >> >[EMAIL PROTECTED] > > >>--------------------------------------------------------------------- >>To unsubscribe, e-mail: jmeter-user- >> >> >[EMAIL PROTECTED] > > >>For additional commands, e-mail: jmeter-user- >> >> >[EMAIL PROTECTED] > > > > > > >-- >Michael Stover >[EMAIL PROTECTED] >Yahoo IM: mstover_ya >ICQ: 152975688 >AIM: mstover777 > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [EMAIL PROTECTED] >For additional commands, e-mail: [EMAIL PROTECTED] > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [EMAIL PROTECTED] >For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
