IBM®
InfoSphere® Big Match for Hadoop version
11.5 supports Hortonworks Data Platform 2.3 and IBM Open Platform/BigInsights
version 4.1.
Procedure
- Export the deployed project from the console and save them
as zip files.
- Run the pme_disable command to disable the tables.
- Uninstall Big Match:
- Follow the instructions in Uninstalling the Big Match Ambari service from your cluster (V11.5.0.5 and above).
- Then, follow the instructions in Removing the Big Match service from Ambari (V11.5.0.5 to V11.5.0.10).
- Remove the old BIGMATCH.repo from
the /var/www/html/repos/* and /etc/yum.repos.d/BIGMATCH.repo directories.
- Restart the services.
- Make sure that the record tables and HDFS workspace are
still intact.
- Complete Ambari upgrade. Follow instructions in Getting Ready to Upgrade Ambari and HDP.
- Complete a rolling upgrade using the instructions in Rolling Upgrade HDP 2.2 to 2.3.
- If you previously used Java 1.7, then upgrade to Java 1.7r9 or higher or Java 1.8. Use
the instructions in Change the JDK Version
- Make sure that all services start and that there are no
warnings or errors. Also, make sure that your HBase data is intact
and accessible.
- Make sure new Big Match prerequisites
are met.
For example, make sure that spark is installed.
- Install the new rpm of Big Match by
using the installation instructions.
- Restart the ambari-server and make sure that httpd is running.
- Install a new version of Big Match by
using the installation instructions.
- Run the pme_enable command on the tables that you disabled
in step 2.
- Import the previously exported project.
- Deploy the project.