Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

When connecting to Argon with FastX version 2 (or later) to open a graphical desktop session such as MATE or KDE, you might get an error like 'The command , such as:

Code Block
The command "mate-session"

...

 exited

or:

Code Block
Could not connect to session bus: Failed to connect to socket

This can occur if you install Anaconda in your home account and select "yes" for the installer to initialize Anaconda with conda init. That option modifies your ~/.bashrc so Anaconda is conveniently configured every time you log in, but the modification also causes graphical sessions to use dbus software supplied in Anaconda, which prevents them from starting. To solve this problem, first check which version of conda you have by running 'conda --version'.

Info

Note that only new sessions you start after making the following changes will be free of the problem; you should discard any existing FastX sessions.

Prior to 4.6.0, edit your ~/.bashrc file to comment out all the lines in the "conda initialize" section. In the future sessions, when you want to use Anaconda in a at your command prompt or in a job script, use the command required to activate ityou'll first need to initialize conda. For example, to activate a default installation of Anaconda3if your Anaconda3 installation uses its installer's default path, use the following command to make conda available:

Code Block
exec "$($HOME/anaconda3/bin/conda shell.bash hook)"

If your conda version of Anaconda is older or if the above command fails, you can still change your PATH variable in your shell or script to access it:

...

is 4.6.0 or newer, the previous method still works, but a better method is available. Instead of commenting out the "conda initialize" section in your ~/.bashrc file, configure conda not to automatically activate the base environment:

Code Block
conda config --set auto_activate_base false

In future sessions (including jobs), you can use 'conda activate' to load any environment you specify.