Installing and uninstalling Big Match Ambari service for Hortonworks or IBM Open Platform (V11.5.0.4)

You can install InfoSphere® Big Match for Hadoop as an Ambari service.

Important: This section covers installing and uninstalling the Big Match Service for Big Match V11.5.0.4. For information about installing or uninstalling Big Match V11.5.0.5, see Installing and uninstalling Big Match Ambari service for Hortonworks (V11.5.0.5 and above). For information about installing or uninstalling Big Match V11.5.0.3 and earlier, see Installing and uninstalling Big Match Ambari service for Hortonworks or IBM Open Platform (V11.5.0.3 and earlier).

The Big Match service consists of several configuration files and two components: Big Match Console and Big Match Runtimes.

The services uses the following configuration files.
  • bigmatch-config.xml - stored in the database and written to disk at /usr/ibmpacks/current/bigmatch/conf/consoleui/bm-config.xml
  • bigmatch-install-config.xml - stored in the database
  • pme-site.xml - stored in the database and written to disk in the /etc/hadoop/conf directory

See the associated property files for a complete description of properties and default values.

Big Match Console

This component is installed on exactly one node. The default lifecycles are responsible for the following actions.

START
Starts the Big Match WebSphere® Liberty Profile instance and HBase REST service (if not already started).
STOP
Stops the Big Match WebSphere Liberty Profile instance.
INSTALL
  • Checks prerequisites (ensures that WebSphere Liberty Profile ports are not already used).
  • Installs the Big Match Red Hat Package Manager (if not already installed).
  • Installs and configures a WebSphere Liberty Profile instance.
  • Creates the Derby database.
  • Creates directories on HDFS for use by various Big Match processes.
  • Alters the configuration files for services that Big Match depends on, including HDFS, HBASE, ZOOKEEPER, YARN, MAPREDUCE2, OOZIE, and KNOX.
  • Adds the Big Match Oozie workflows and sample property files to HDFS.
  • Adds libraries to the Oozie server WAR file.
CONFIGURE
Does nothing.
STATUS
Returns the status of the WebSphere Liberty Profile instance.
RESTART
Runs STOP followed by START.

This component has two custom commands:

ADD_SAMPLE_DATA
Creates the bmSample1 table in HBase. Adds the sample data load file and a Big Match sample configuration for this table to HDFS. The table is then pme_enabled and loaded with 10,000 sample records.
REMOVE_SAMPLE_DATA
Removes the Big Match sample data load file from HDFS and removes the bmSample1 table from HBase.

Big Match Runtimes

This component is installed on one or more nodes and the default lifecycles are responsible for the following actions.

INSTALL
  • Installs the Big Match Red Hat Package Manager (if not already installed).
  • Creates the Big Match user and group 'bigmatch' if they do not exist.
  • Creates the bigmatch-env.sh file.
  • Configures the node based on what components are installed. If service HBASE has components (HBASE_CLIENT, HBASE_MASTER, HBASE_REGIONSERVER) installed on the node, then the Big Match HBase table extensions are added to the HBASE_CLIENT components lib directory. Additionally the HBase shell ruby script is extended with the pme library and its custom commands. If the node has the HBASE_REGIONSERVER component installed, the Big Match queue directory is created. If service HDFS has components (HDFS_CLIENT, DATANODE) installed on the node, then the Big Match Hadoop extensions are added to the HDFS_CLIENT components lib directory.
CONFIGURE
Writes the pme-site configuration to file /etc/conf/hadoop/pme-site.xml. Writes the bigmatch-config configuration to file /usr/ibmpacks/current/bigmatch/conf/consoleui/bm-config.xml.
Note: The INSTALL and CONFIGURE lifecycles run each time an Ambari client component is restarted.