MTT

Middleware Testing Tool

Getting Started With MTT

MTT can be cloned from https://github.com/open-mpi/mtt.

The path to your MTT directory needs to be added to the the environment. Add
export MTT_HOME=/path/to/mtt
to your .bashrc or .bash_profile and activate the file using ‘source .bashrc’ or ‘source .bash_profile’ at the command line.

When you are finished with this guide head on over to the Plugins and INI guides to get familiar with them.

Set Up a Python Virtual Environment

When you first try to run the MTT program you may get the error:

File “pyclient/pymtt.py”, line XX, in from SOMETHING import SOMETHING ImportError: No module named SOMETHING

Where SOMETHING represents some missing python package. This indicates that you need to set up a Python virtual environment to run MTT in.

Python virtual environments create an isolated environment for python based applications. Using a virtual environment for MTT means you can install all necessary dependencies without worrying about affecting other applications. This document covers two ways to set up python virtual environments. The Conda package and environment manager is the easiest to use. It is included with all versions of Anaconda and Miniconda. If you do not have access to Conda then you will need to install the virtualenv and virtualenvwrapper packages.

The virtual environment only has to be set up once and then it can be called anytime the user needs it. If installing on a system with front ends and compute nodes the virtual environment should be set up on the front end only.

Environment Setup With Conda

Set Up ~/.condarc

Before creating a conda environment, a user will need to set up a .condarc file that will tell conda where to put conda environments.
Default environment directories are:
$HOME/anaconda2/envs
$HOME/.conda/envs

To set up your .condarc to use a different location to store conda environments issue the following commands.

$ conda config --add envs_dirs /path/to/your/envs

This will create ~./condarc if it doesn’t exist and will add the necessary information to it.

Verify That ~/.condarc is Correctly Set Up

Using the ‘conda info’ command:


$ conda info  
     active environment : None  
       user config file : /Users/username/.condarc  
 populated config files : /Users/username/.condarc  
          conda version : 4.5.11  
    conda-build version : 3.10.5  
         python version : 2.7.15.final.0  
       base environment : /Users/username/anaconda2  (writable)  
           channel URLs : https://repo.anaconda.com/pkgs/main/osx-64  
                          https://repo.anaconda.com/pkgs/r/noarch  
                          https://repo.anaconda.com/pkgs/pro/osx-64  
                          https://repo.anaconda.com/pkgs/pro/noarch  
          package cache : /Users/username/anaconda2/pkgs  
                          /Users/username/.conda/pkgs  
       envs directories : /Users/username/anaconda2/envs  
                          /Users/username/.conda/envs  
               platform : osx-64  
             user-agent : conda/4.5.11 requests/2.18.4 CPython/2.7.15 Darwin/17.7.0 OSX/10.13.6  
                UID:GID : 502:20  
             netrc file : None  
           offline mode : False

Check the ‘envs directories:’ it should be set to the directory given to the ‘conda config –add’ command. Redo if necessary (remove bad values by using ‘conda config –remove envs_dirs /path/to/remove or by editing ~/.condarc)

NOTE: Conda version 4.5.12 is putting the newly created environment into the current working directory so you may need to launch the conda env create command from the directory you desire it to end up in.

Create Your Conda Environment

Use the condaEnv.yml file in the mtt directory as follows:

$ conda env create -f /path/to/mtt/condaEnv.yml

The default name for the environment is MTTEnv. You can change it if you wish by editing the condaEnv.yml file ‘name:’ section.

Enter or Exit the Conda Environment

When the conda environment is active <MTTEnv> will be the first part of your terminal descriptor. To activate a environment use ‘source activate MTTEnv’ or ‘conda activate MTTEnv’. To exit an environmnt use ‘source deactivate’ or ‘conda deactivate’.

Some other usefull conda commands are:
‘conda list’ to see a list of all the packages in the current environment.
‘conda env remove <name>’ to remove an environment.
‘conda install <package name>’ to individually install packages to the current environment.
‘conda remove –name <Environment name> <Package name>’ to remove a package from an environment.

More information is available at docs.conda.

Environment Setup With Virtualenvwrapper

Install virtualenv and virtualenvwrapper as follows:

$ pip install virtualenv
$ pip install virtualenvwrapper

Edit ~/.bashrc or ~/.bash_profile

Before creating a python virtual environment with virtualenvwrapper confirm the path to virtualenvwrapper.sh. By default it should be in either:
/usr/local/bin OR
$HOME/.local/bin

Add the following lines to your ~/.bashrc file or equivalent.

source /path/to/virtualenvwrapper.sh
export WORKON_HOME=$HOME/virtualenvs

Refresh your bash shell.

$ source ~/.bashrc  (or .bash_profile)

The script should run creating several files needed by virtualenv.

Create Your Virtual Environment

Use the pyenv.txt file in the mtt directory as follows:

$ mkvirtualenv -r /Path/to/mtt/pyenv.txt <env_name>

Enter or Exit the Virtual Environment

To activate the environment use:

$ workon <env_name>

To exit the virtual environment type:

$ deactivate

More information is available at virtualenvwrapper.

Problems you may encounter.