Restrictions for upgrading R and Python packages
Some R and Python packages shall not be upgraded because doing so will break the Workbench UI.
For R Runtimes
Do not upgrade the Cairo
or RServe
packages in your
projects using R Runtimes. If you do, you may be unable to launch sessions, jobs, experiments,
or applications.
If these packages are upgraded, you will need to start a new project or to delete the
directory .local/lib
using the CDSW/Cloudera AI Files UI ("show hidden").
For Python Runtimes
Do not upgrade the ipykernel
package or its dependencies
(ipython
, traitlets
, jupyter_client
, and
tornado
) in your project using Python Runtimes. If you do, you may be
unable to launch sessions, jobs, experiments, or applications.
If these packages are upgraded, you will need to start a new project or to delete the
directory .local/lib
using the CDSW/Cloudera AI Files UI ("show hidden").
Python Runtimes in Cloudera AI fail to import the setuptools
Python library and
can fail installing some Python packages
Python Runtimes in Cloudera AI fail to import the setuptools
Python library and
therefore can fail installing some Python packages when the library
setuptools
is present on the Runtime or is installed into the Cloudera AI project
with version 60.0.0 or higher.
Python 3.10 Runtimes from the 2023.05 Runtime release ship with a newer version of
setuptools
, so customers can run into this issue when they are using that
Runtime. Also they can run into this issue when they are using Custom Runtimes that has a
newer setuptools
library version or when they install a new
setuptools
version into their project (regardless of what Runtime they
use).
Workaround: Set the environmental variable
SETUPTOOLS_USE_DISTUTILS=stdlib
either on a project level under
Project Settings -> Advanced or on a workbench level under
Site Administration -> Runtime -> Environment variables.