Using publishlibs
Command Line Usage: publishlibs
The publishlibs script is used to create an archive of the installed DataFlow libraries and publish the archive to HDFS in the /apps/actian/dataflow/archive directory. The DataFlow libraries are found in the /opt/datarush/lib directory by default. The full contents of the lib directory are archived into a tar.gz (gzipped tar file) and cached in the local /opt/datarush/archive directory.
The publishlibs script is executed by the clustermgr script when the start-history-server command is used to start the cluster manager for YARN integration. The publishlibs script can also be run directly from the command line as needed. To refresh the library archive contents in HDFS, simply run publishlibs. The script will always overwrite the contents in HDFS to ensure the latest version of the DataFlow libraries are contained in the library archive.
Note: The /apps/actian/dataflow directory must already exist in HDFS and be owned by the actian user. Please update the variable HDFS_TARGET_DIR to modify the target directory in HDFS where the DataFlow library archive will be pushed.
The command line syntax is:
publishlibs [ archive_path ]
Last modified date: 12/09/2024