Submit your use case or success story to the 2023 edition of the Dataiku Frontrunner Awards ENTER YOUR SUBMISSION

Kernel not found

Solved!
dot101
Level 2
Kernel not found

So when i run the R recipe i get this error :

Job failed: environment variable DKURBIN not defined

 

And when i open the code in the notebook it shows this :

Kernel not found

Could not find a kernel matching R. Please select a kernel

 

 

How can i solve this ?

Thanks in advance.

0 Kudos
1 Solution
sergeyd
Dataiker

Hi @dot101 

This most probably means that your R subsystem is installed in a place where DSS cannot find it (in particular: not in the system's default search PATH for commands)
This is unfortunately something that is known to happen on MacOS, especially when you start DSS with the Finder app.
Details of this issue would be:
- you have installed R somewhere on the system. On MacOS, there are various possible sources for this, with various installation locations (eg /usr/local/bin, /opt/local/bin, etc)
- somehow, this install location was added to your command-line session initialization file ($HOME/.profile, $HOME/.bash_profile or equivalent) so that it is added in the PATH environment variable upon command-line shell startup
- as a consequence, if you run the DSS installer, or "dssadmin" post-install command, or even start DSS from a shell window, it inherits this PATH setting and can find R
- on the contrary, if you start DSS from the Finder app, the shell initialization profile files are not loaded, this PATH entry is not defined, and DSS cannot find R anymore
You can fix that with the following:
- open a shell session, and look up the install location for R with: "which R"
- you can even test that if you restart DSS from this command line session, as in:
    /Users/<user>/Library/DataScienceStudio/dss_home/bin/dss restart
  it should now find R and be able to launch R recipes and notebooks
- you can force DSS to use a given location for R by defining the following in file <DSS_DIR>/bin/env-site.sh :
  # Adjust this to point to the result of "which R" above
  export DKURBIN=<result_of_which_R_command>
- you need to restart DSS to take this modification into account

View solution in original post

1 Reply
sergeyd
Dataiker

Hi @dot101 

This most probably means that your R subsystem is installed in a place where DSS cannot find it (in particular: not in the system's default search PATH for commands)
This is unfortunately something that is known to happen on MacOS, especially when you start DSS with the Finder app.
Details of this issue would be:
- you have installed R somewhere on the system. On MacOS, there are various possible sources for this, with various installation locations (eg /usr/local/bin, /opt/local/bin, etc)
- somehow, this install location was added to your command-line session initialization file ($HOME/.profile, $HOME/.bash_profile or equivalent) so that it is added in the PATH environment variable upon command-line shell startup
- as a consequence, if you run the DSS installer, or "dssadmin" post-install command, or even start DSS from a shell window, it inherits this PATH setting and can find R
- on the contrary, if you start DSS from the Finder app, the shell initialization profile files are not loaded, this PATH entry is not defined, and DSS cannot find R anymore
You can fix that with the following:
- open a shell session, and look up the install location for R with: "which R"
- you can even test that if you restart DSS from this command line session, as in:
    /Users/<user>/Library/DataScienceStudio/dss_home/bin/dss restart
  it should now find R and be able to launch R recipes and notebooks
- you can force DSS to use a given location for R by defining the following in file <DSS_DIR>/bin/env-site.sh :
  # Adjust this to point to the result of "which R" above
  export DKURBIN=<result_of_which_R_command>
- you need to restart DSS to take this modification into account