Standalone Upgrade
{% hint style="success" %} Before proceeding with any upgrades, please remember to backup your DQ Metastore {% endhint %}
{% hint style="danger" %} Please remember that rolling back Collibra DQ to a prior version is not supported. Please contact Collibra Support with any questions. {% endhint %}
Download DQ Upgrade Package
Note: Beginning December 2021, all Collibra DQ customers upgrading or patching will be receiving the Full package (vs. Base) and should follow the same Upgrade steps below
Download tarball using the signed link to the full package tarball provided by Collibra. Replace <signed-link-to-full-package> with the link provided.
### Go to the OWL_BASE (home directory of the user is most common)
### This example we will use /home/owldq installing as the user owldq
cd /home/owldq
### Download & untar
curl -o dq-full-package.tar.gz "<signed-link-to-full-package>"
tar -xvf dq-full-package.tar.gz
### Clean-up unnecessary tarball (optional)
rm dq-full-package.tar.gz
Upgrade Steps
- Copy the contents of the provided package e.g. owl-<newversion>-<SPARK301>-package-full.tar.gz to the system being upgraded (extract contents)
- Best practice: Untar contents into a uniquely named folder e.g. 2021-12-dq-upgrade
- Stop owlweb using
./owlmanage.sh stop=owlweb - Stop owlagent using
./owlmanage.sh stop=owlagent - Move old jars from
owl/binmv owl-webapp-<oldversion>-<spark301>.jar /tmpmv owl-agent-<oldversion>-<spark301>.jar /tmpmv owl-core-<oldversion>-<spark301>.jar /tmp
- Copy new jars into the
owl/binfolder from the extracted packagemv owl-webapp-<newversion>-<spark301>.jar /home/owldq/owl/binmv owl-agent-<newversion>-<spark301>.jar /home/owldq/owl/binmv owl-core-<newversion>-<spark301>.jar /home/owldq/owl/bin
- run
./owlmanage.sh start=owlwebto start the owl-web application - run
./owlmanage.sh start=owlagentto start owlagent
Additional Notes / Steps Due To Log4J (December 2021)
Additional Step 1: Place Log4j-1.2-api-2.17.1.jar (as of 2022.02) into /<install-home>/owl/spark/jars
-
Note: Was Log4j-1.2-api-2.17.0.jar in 2021.12 and 2022.01
Who: All Collibra DQ customers, particularly those leveraging CLI mode
- Navigate to the same folder where the Collibra provided upgrade package was extracted
- Navigate to <location of 2022-02-dq-upgrade>/packages/install-packages
- Extract the needed log4j-1.2-api-2.17.1.jar via the command:
tar -xvf spark-extras.tar.gz spark-extras/log4j-1.2-api-2.17.1.jar - Move the log4j-1.2-api-2.17.1.jar file into /<install-path>/spark/jars folder
FAQ
Q: (When) do I need to move Log4j-1.2-api-2.17.1.jar before or after swapping the main Collibra DQ jars?
- A: Sequence does not matter.
Q: (What) if I don't follow these additional upgrade steps?
- A: If your
SPARK_SUBMIT_MODEwithin owl-env.sh is set toSPARK_SUBMIT_MODE=native, Collibra DQ will function properly without the above additional upgrade step, with the exception of CLI mode
Additional Step 2: Remove a legacy properties file
Who: Only Collibra DQ customers upgrading Agents installed on Cloudera CDP Hadoop Edge Nodes
- Navigate to /<agenthome>/owl/config/
- Remove
log4j-cluster.propertiesfile
FAQ
Q: (When) do I need to remove log4j-cluster.properties before or after swapping the main Collibra DQ jars?
- A: Remove the file before restarting owl-agent. Otherwise, stop owl-agent again, remove the file, then restart owl-agent.
Q: (What) if I don't follow these additional steps?
- A: For customers using agents on Hadoop edge nodes, they will receive errors when running DQ Jobs due to engaging a method that no longer exists.
Q: What should I do if I am not a vendor-supported Cloudera CDP version?
- A: Our testing and guidance mainly applies to vendor-supported (non-EOL) Cloudera CDP versions. Other Hadoop variants may handle logging differently and may require the legacy properties file. In short, feel free to first upgrade without this step, then remove the log4j-cluster.properties file if DQ Jobs are running into issues.