Oracle Cloud Infrastructure Documentation

Ambari Services Log Paths

Some long-running clusters with high workloads generate high volumes of log content, which might fill most of the disk space with log files or tracing files in the boot volume. If the OS boot volume is full, it might also impact the cluster's health. Therefore, Big Data Service moved the log directory from the community default path (/var/log) to a block volume mounted directory (/bdslogs) in new clusters. If you use /var/log, or any other path, you can customize it through ODH Utilities.

Note

The path selected is the default path for new services in Ambari and other tracing files such as crash dumps (.hprof) and HeapDump files.

Customizing the Common Log Path

You can customize the Big Data Service common log path.

Note

During ODH patching, from old clusters that use /var/log, the default log path isn't changed to avoid confusion or incompatibility of any tools that might use this path. We recommend you manually update the path to /bdslogs.

Ambari-server and ambari-agent require a manual restart when the log directory is changed.

  1. Access Apache Ambari.
  2. From the side toolbar, under Services click ODH Utilities.
  3. Click the Configs tab, and then select ODH LOGGER CONFIG.
  4. Update Log Directory configuration value.
    Note

    The log directory is the base directory where all ODH service daemon logs are written. For example, /bdslogs/hadoop-hdfs/, /bdslogs/yarn/, /bdslogs/spark3. The default value is /bdslogs.
  5. Click Save.
  6. Click the Services '...' icon, and then click Restart all required.