Frequently asked questions about Multi-Mechanize

What is the minimum version of Python I should run?

2.6 is required
2.6.3+ is reccomended (http://bugs.python.org/issue6838)

Which operating systems are supported?

It should run anywhere you can install Python and Matplotlib. Development and testing is mostly done on Ubuntu 64-bit and Windows 7 64-bit.

I have received reports of people having trouble getting Matplotlib installed on OS X, but don't have an environment to confirm this on.

Why does it launch multiple OS processes?

Each "user group" runs in its own process. Within each process, multiple virtual user agents are run; each in its own thread. So it uses a hybrid multi-process/multi-threaded architecture.

How do you record HTTP traffic?

There is no "recorder" supplied with multi-mechanize. You have to write the scripts yourself. Script examples are shipped with the tool to get you started.

There are also several browser plugins that will help you view HTTP traffic and see what is going on at the network layers:

• Fiddler
• Firebug
• HTTPFox

How do I install Mechanize on Windows?

• go to http://pypi.python.org/pypi/distribute#distribute-setup-py
• download the distribute-setup-py script
• run the distribute-setup-py script. this will install Distribute
• go to a command prompt at C:\Python26\Scripts
• run: 'easy_install mechanize'

Where do I get Matplotlib and Numpy for 64 bit windows?

If you are running Windows 64-bit, I recommend installing Numpy then Matplotlib from here:

• http://www.lfd.uci.edu/~gohlke/pythonlibs/

Which packages do I need to install on Ubuntu/Debian?

If you are running Ubuntu or Debian, you can get all the prerequisites directly from the OS package repository. Run the following commands:

• sudo apt-get install python-mechanize
• sudo apt-get install python-matplotlib

What are the fields in results.csv?

results.csv is used for results storage during a test run. It is a comma separated text file that is used internally to generate results reports. You might also find the raw file useful for your own analysis.

the fields correspond to:
transaction count, elapsed time, epoch time, user_group name, script run time, error, custom_timers

Should I keep console_logging on during my tests?

No, only keep logging on while developing/debugging your scripts. You should turn it off to get highest throughput. It takes substantial system resources to print all responses to stdout when doing high amounts of i/o.