ciso8601 library stopped installing out of the blue on DB clusters

[RasmusOlesen]

We have multiple DB clusters (6.4 Extended Support) that have not changed in terms of libs installed or nodes etc.

Sudden from one day to the other, after a cluster restart August 7th, they stopped installing ciso8601 lib as they would usually.

Anybody know what would cause this?

We have literally changed nothing on the clusters.

[SarahPedersen]

We also have 6.4 cluster and our ciso doesn't work either.

[RasmusOlesen]

Output from the lib installation under the Clusters page (Pypi: ciso8601)

java.lang.RuntimeException: ManagedLibraryInstallFailed: org.apache.spark.SparkException: Process List(/databricks/python/bin/pip, install, ciso8601, --disable-pip-version-check) exited with code 1. Command "/databricks/python3/bin/python3.7 /databricks/python3/lib/python3.7/site-packages/pip/_vendor/pep517/_in_process.py get_requires_for_build_wheel /tmp/tmpp25i_3y_" failed with error code 1 in /tmp/pip-install-xva3elqs/ciso8601
 for library:PythonPyPiPkgId(ciso8601,None,None,List()),isSharedLibrary=false

[SarahPedersen]

We have the same problem. We're also running on a 6.4 cluster and get the same error

[siddhathPanchal · Accepted solution]

Hey @Sarah Pedersen​ and @Rasmus Olesen​ . I am able to download that library successfully on my DBR 6.4 cluster. Please look at the SS attached. If you are using any outdated version of that library, I would highly suggest to use the latest version of that library and install it from the cluster UI.

[RasmusOlesen]

Just to close this old qustion:

We solved this by switching to a PEP517 free pip install, using the a Global Init Script:

/databricks/python/bin/pip install ciso8601 --disable-pip-version-check --no-use-pep517

Now it works for us.