Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Table of Contents
The following release notes cover the version of Qlik Catalog released in May 2023. For questions or comments, post in the Product Forums or contact Qlik Support.
The trailing slash in the /isUserNameAvailable endpoint caused unexpected encoding results in some cases. Removing it resolved the encoding issue, providing consistent results for endpoint invocations.
The code that manages permission changes over published files didn’t support whitespaces in entity names. This fix enables support for such cases.
During the Catalog Installer run on Ubuntu OS, pulling the pause-3.5 container required by podman sometimes failed. To address this, the pause-3.5 container is now included with other containers in the QDC binaries.
A classpath issue was identified during interactions with a Kerberized Hadoop/Hive cluster, where a mix between the isolated driver classpath and the general classpath caused library conflicts, preventing communication with the cluster. This fix ensures that the Hive JDBC driver is fully isolated from the main application classpath.
Two example shell scripts have been added to the Swagger documentation:
On the Source module, "Audit Logs" tab, an audit entry is created for each load, prepare and publish job that is executed, whether run immediately or scheduled.
However, the "User Name" was incorrectly recorded as "ANONYMOUS" rather than the user that initiated the job. This has been fixed.
Please note that the "IP" and "Browser Information" fields were also set to "ANONYMOUS". These two fields are now set to "NOT_APPLICABLE".
For customers that have previously run thousands of load jobs, each of which may have large details, query performance when displaying a Source's entities, or an entity's load logs (and details), could drastically decrease, even resulting in OutOfMemory. One solution is to delete old load logs. Alternatively, a new core_env property has been introduced to alter how the backend queries load logs and their details. Consider setting this property to false if this issue arises:
# There are entity retrieval APIs (e.g., /entity/external/entitiesBySrc/{srcId})
# that query the database for all load logs (aka work orders) for the entities.
# For tens-of-thousands of work orders, with large details, this can overwhelm
# the backend, resulting in OutOfMemory. Setting this property to false will
# skip the work order database query. To enable this workaround, set this
# property to false. NOTE: If set to false, total and latest record counts will
# display as "N/A" in the Source module entity grid table. The record counts are
# available in the Discover module entity grid table or when viewing load logs
# (at either the grid table or detail level).
#always.attach.load.logs.and.details=true
Two improvements were made to Catalog installation on RHEL 9.2
Two enhancements were made to Catalog's support for SAML single sign-on (SSO)
A sample bash shell script is now included to demonstrate how to use the API, via curl, to initiate an entity load. Please see file /usr/local/qdc/apache-tomcat-9.0.85/webapps/qdc/resources/api/loadDataForEntities.sh
One of the features of Catalog is its detection of metadata changes, particularly for JDBC sources. For REGISTERED and ADDRESSED entities, Catalog detects column additions and deletions, as well as column type changes. This capability is controlled by the following core_env property:
# If true, for:
# (1) ADDRESSED and REGISTERED entities loaded from JDBC relational sources, and # (2) ADDRESSED entities loaded from FILE sources, then
# column additions and deletions will be detected and applied. If true, business metadata, tags and properties will be
# preserved and not overwritten. Set to false for legacy behavior. Default: false (default to false so it is disabled
# for legacy customers that upgrade)
support.schema.change.detection=true
However, when this property was set to "false", for ADDRESSED entities, Catalog was detecting and applying column additions and deletions when it should not have done so.
With this fix, if the above property is "false", for ADDRESSED entities, Catalog will only detect column type updates. This change enables customers initially to select a subset of a table's columns and to have that subset remain unchanged as entities are later loaded.
When installing on Ubuntu 22.04, the loading of the engine and licenses images did not result in the "pause" image being loaded. This prevented those containers from starting -- the following error was seen:
$ ./launch_qlikContainers.sh
Launch Containers with Podman
Error: k8s.gcr.io/pause:3.5: image not known
The installer has been modified to check for this missing image, and to pull it if necessary.
Audit Log entries are now created for the following changes (including additions and deletions) to Sources, Entities, and Fields:
business metadata
properties
tags
This feature has been implemented and tested for two common scenarios:
import of business metadata
manual edit in Catalog user interface (UI)
The ordering and display of Audit Log columns has changed:
The Catalog UI exhausted web browser memory when viewing the Admin / Import/Export Metadata page if there were many completed job histories and if the details of given job histories were megabytes in size. This has been addressed by truncating the details of each import/export job when the entire page of job histories is retrieved. The full details are still returned when "View Details" is invoked for an individual record.
Standard upgrades of vulnerable third-party dependencies were made.
Standard upgrade of Apache Tomcat (for first-time installs) was made.
Data load jobs may now be scheduled for automatic execution. To schedule a recurring data load, visit the Source module, select an entity, and initiate a load. Click on the Scheduling expander in the Data Load modal and enter a Quartz expression, such as:
Once the Quartz cron expression has been entered, click OK to schedule the data load job.
For more information, please visit the online help and search for "Scheduling" -- the topic is "Qlik Catalog Data ingest: Loading data".
Qlik Catalog no longer offers a multi-node deployment option. Multi-node deployment allowed Catalog to manage a Hadoop cluster data lake, where data was stored in HDFS or S3 and queried using Hive. Most recently, Catalog supported the Cloudera Data Platform (CDP) and AWS EMR platforms. Consequently, the installer property INSTALL_TYPE is no longer present and need not be set -- all installations now use the single node deployment option.
Standard upgrades of vulnerable third-party dependencies were made.
Standard upgrade of Apache Tomcat (for first-time installs) was made.
Jira ID: QDCB-553
Source and entity properties were editable in the case where the "editable" attribute of the property definition was set to false. This has been fixed. Note that these properties may still be deleted (removed) from the entity or source. In the following screenshot, the property "scheduled.load.job.configuration" is not editable:
Download this release from the Product Downloads page on Qlik Community.
About Qlik
Qlik’s vision is a data-literate world, where everyone can use data and analytics to improve decision-making and solve their most challenging problems. A private SaaS company, Qlik offers an Active Intelligence platform, delivering end-to-end, real-time data integration and analytics cloud solutions to close the gaps between data, insights and action. By transforming data into Active Intelligence, businesses can drive better decisions, improve revenue and profitability, and optimize customer relationships. Qlik does business in more than 100 countries and serves over 50,000 customers around the world.