Fixed issues in Cloudera AI on premises 1.5.5

This update standardises the use of absolute timestamps across the entire application, replacing relative timestamps such as 1 week ago with precise date and time information. The Files section in the Project Overview page now displays exact Last Modified dates, while the Jobs section in the Job Overview page shows accurate Latest Run times for improved clarity and consistency.

The Livelog cleaner previously cleaned up input and output log topics but did not delete the content of container log topics. It now includes engine containers and Spark executor log topics in the cleanup process.

Previously, Spark workloads were unable to read the ZSTD-encoded Hive tables when running in the default Spark on Kubernetes mode.

This issue has now been resolved, enabling Spark to successfully read ZSTD-encoded data from Hive tables.

The Shared Memory configuration set in Project Settings > Advanced is now applied to applications in addition to sessions.

This configuration is retrieved from the Projects table when a user creates a new application.

The Experiment page on the Details tab occasionally displayed a negative value for the duration of time, due to an incorrect timezone configuration in certain setups. This issue has been solved.

The following error was intermittently observed in the db-0 pod during restarts or upgrades:

• CDSW_PG FATAL: lock file "postmaster.pid" already exists

• CDSW_PG HINT: Is another postmaster (PID 7) running in data
          directory

This issue occurred because the db-migrate-xxx job was still in the Running state and was attempting to access the embedded database.

The issue has now been resolved. The db-0 pod no longer enters a crash loop due to a stale postmaster.pid file.

This update addressed an issue where the Workbench editor occasionally opened the wrong file. The problem stemmed from timing conflicts in DocumentManager.js, where a rapid refresh could retrieve an outdated file from local storage instead of the file that was just clicked.

The fix ensures that the DocumentManager.js consistently selects the most recently clicked file during the refresh process, even in scenarios involving quick refreshes.

There was an issue in the Workbench editor where the Find input box (Ctrl+F or Command+F) retained the previous search string upon reopening. The input box now clears automatically when clicking anywhere in the editor text area, enabling you to enter a new search string seamlessly.

Earlier, the HTML code generated to embed an image in PBJ Workbench did not work. Now, you can embed images generated in PBJ Workbench-based sessions similarly to how you can embed images from Workbench-based sessions using the share icon next to the generated images.

PBJ Workbench Runtime images now comply with the value set in the environmental variable MAX_TEXT_LENGTH. This limits the maximum number of characters that can be displayed by each command executed.

Earlier, the Interrupt button did not work in Cloudera AI sessions that ran a PBJ Workbench Runtime. This issue has been fixed.

Sessions using PBJ-based custom Runtimes with a custom editor could not start previously. This issue has been fixed.

When PBJ Workbench R Runtime was used, the tables and help text were not properly displayed. This issue has been fixed.

When PBJ Workbench R Runtime was used, the help text, required as a session command's output, was not properly displayed. This issue has been fixed.

When PBJ Workbench R Runtime was used, the content of the tables was not properly displayed. This issue has been fixed.

When PBJ Workbench R Runtime was used, the start of each command execution was logged with INFO severity, and was consequently displayed. Now, the command execution is logged with DEBUG severity and is not displayed.

When PBJ Workbench R Runtime was used, the editor was not able to show code completion. This issue has been fixed.

A previously unauthenticated endpoint is now securely authenticated through a reverse proxy, using a mechanism called Browser Accessible Service.

PBJ Workbench jobs now respect the JOB_MAXIMUM_MINUTES environment variable. PBJ Workbench jobs are automatically terminated once they exceed the time limit specified by JOB_MAXIMUM_MINUTES.

When SESSION_MAXIMUM_MINUTES is set, sessions are expected to expire after the specified duration, regardless of activity (default is 7 days). However, unlike PBJ Sessions and Remote Editor Sessions, Jupyter Sessions previously continued running beyond this time and did not time out. This issue has now been resolved.

This update improves the UI on both the Job Creation page and the Job Settings page to clarify that the PBJ-based ML Runtime transforms job arguments into the JOB_ARGUMENTS environment variable. It introduces a dedicated label and a help block, including a link to a code snippet for user guidance.

Project creation using a private repository with SSH key and SSH URL did not work on NTP setups. This issue has been fixed.

Previously there has been an error fetching the AMP catalog on NTP/Airgap Proxy Openshift setup when selecting Site Administration > AMP tab.

This issue has been solved.