Troubleshooting the Adapter

Note: If you have updated the passwords and are experiencing issues with running jobs or events, we recommend disabling and re-enabling the connection to ensure everything works properly.

Review Service Log Files for More Information

Refer to the log files for further information regarding an issue.

Connection Failures

  • Verify the hostname to IP address mappings in the hosts file.

  • Verify service.props has been correctly set up to support Kerberos for secured connections.

  • Verify that the user keytab file exists and is accessible to TA master.

  • Verify that the MapR client is configured correctly.

Hadoop:GSSException: No valid credentials provided :Clock skew too great

Occurs when the clock skew between the KDC and clients exceed a maximum threshold (default 5 minutes). Maintaining synchronized clocks between the KDCs and Kerberos clients (TIDAL) is required, therefore Network Time Protocol (NTP) software or other similar time service tools must be used to synchronize them. Updating the Master's clock to the KDC server time will temporarily address the issue, but a time service software must be used to keep them synchronized.

Job Failures

  • Verify your job is configured correctly.

  • Verify your job can be run via the Hadoop CLI before running thru TA .

  • Check the Adapter logs and verify how your job ran from the Hadoop Admin Console.

  • Verify the file paths and names are case-sensitive and that they exist on the HDFS.

  • Verify that the job JAR, as well as all referenced library JARs, exists in HDFS and that the path specified is correct.

Adapter Is Out-of-Memory

Adapter memory sizes are verified on a 10-node cluster and can be increased in the Adapter service.props.

Output Files Cannot Be Viewed

The output file either contains binary data or its size is greater than default size 1MB.

MapReduce Job Runs Fine on the Hadoop Client but Fails via TA Adapter

Most likely, your TA MapReduce job configuration is not setup correctly. Compare the Hadoop job configuration file on the cluster with TA to resolve the problem.