How to Clear Hadoop/HDFS Cache for Infoworks Jars

Issue

Clearing Hadoop/HDFS cache for jars packaged from Infoworks.

Solution

  • Login to the Infoworks edge node (the machine where Infoworks is installed).
  • Run the below commands:

cd $IW_HOME/bin

source env.sh

./stop.sh hangman

cd $IW_HOME/conf

cat conf.properties | grep hdfs_temp

NOTE: Ensure that no Infoworks jobs are running. You can verify the job queue in the Infoworks ADE Admin page.

  • Copy the iw_hdfs_temp_home path displayed in the output. For example, iw_hdfs_temp_home=/user/ec2-user/temp.
  • Remove the jars and path from the specified directory using the output from above command as base path.

For example,

hadoop fs -rm -R output_of_step_3 / iwjobs/libjars/opt/infoworks/

For Ingestion and related jars,

hadoop fs -rm -R output_of_step_3 /udfs/opt/infoworks

For UDF and related jars,

hadoop fs -rm /user/ec2-user/temp/iwjobs/libjars/opt/infoworks/

hadoop fs -rm /user/ec2-user/temp/udfs/opt/infoworks/

  • Run the following commands:

cd $IW_HOME/bin

sourcce env.sh

./start.sh hangman

  • Login to the Infoworks ADE and rerun the job.
Type to search, ESC to discard
Type to search, ESC to discard
Type to search, ESC to discard