Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
We have been trying to update some library versions by uninstalling the old versions and installing new ones. However, the old libraries continue to get installed on cluster startup despite not showing up in the "libraries" tab of the cluster page. We can see in the log4j logs that they're getting installed, and we can import them in notebooks to verify that their versions are what we're trying to uninstall. Is this a known issue? How can I view/modify what the cluster decides to install on startup if it's not the libraries tab?
The issue seemed to go away on its own. At some point the libraries page started showing what was getting installed to the cluster, and removing libraries from the page caused them to stop getting installed on cluster startup. I'm guessing there was some bug causing a disconnect between what the libraries page was showing and what the cluster was actually using to install libraries. I doubt it was anything specific to that cluster (like init scripts, etc.) as we were seeing it on other clusters too.
could you please check if there are any global init scripts specified at the workspace level? or init scripts specified in the cluster from which the libraries are getting installed.
Hi @Reed Zhang , This article describes how to manage Azure Databricks clusters, including displaying, editing, starting, terminating, deleting, controlling access, and monitoring performance and logs.
Please go to the library and check if this cluster is marked to get the library automatically installed on the cluster. If so please uncheck it to resolve the issue.
The issue seemed to go away on its own. At some point the libraries page started showing what was getting installed to the cluster, and removing libraries from the page caused them to stop getting installed on cluster startup. I'm guessing there was some bug causing a disconnect between what the libraries page was showing and what the cluster was actually using to install libraries. I doubt it was anything specific to that cluster (like init scripts, etc.) as we were seeing it on other clusters too.
Hi @Reed Zhang , Just a friendly follow-up. Do you still need help, or have you helped yourself to find the solution? Please let us know.
Connect with Databricks Users in Your Area
Join a Regional User Group to connect with local Databricks users. Events will be happening in your city, and you won’t want to miss the chance to attend and share knowledge.
If there isn’t a group near you, start one and help create a community that brings people together.
