DataTap Issues

If creating a DataTap fails, you can view the following logs on the primary Controller to diagnose the issue:

/var/log/bluedata/bds-mgmt.log
/var/log/bluedata/bds-dataserver.log
/var/log/bluedata/bds-cachingnode.log

For more information about the logs, see Platform Logs.

Symptom Troubleshooting/Resolution
Server <Kerberos TGT> not found in Kerberos database. This may be due to a Kerberos configuration issue. Collect and analyze /var/log/bluedata/bds-cachingnode.log for information about the problem.
Container is reporting an HDFS error when running a Kerberized DataTap.
  1. Determine whether the DataTap is configured for proxy or passthrough mode.
  2. Collect and analyze /var/log/bluedata/bds-cachingnode.log from the physical host where the container resides.
  3. If the cnode log indicates a connect() error, then collect the following:
    • core-site.xml and hdfs-site.xml on the remote HDFS.
    • Get the Datanode log file.
    • Verify that the Datanode is listening on the port 1004 for non-Kerberized or on port 50010 for Kerberized DataTap by executing the command netstat -tulnp | grep <port_num>, where <port_num> is the port number.
    • Verify that the HPE Ezmeral Runtime Enterprise Worker host can ping the DataNode and vice-versa.
    • Copy the file to the remote HDFS directly by executing the command hdfs://namenode:port/path in the Kerberized compute cluster to bypass the HPE Ezmeral Runtime Enterprise Java client code.