HST agents collect data for the specific node on which they are installed. To capture data for all nodes in the cluster, which is the most common use case, you must run the hst capture command on all nodes. Typically this is done using pdsh or other parallel distributed shell utilities. Running the hst capture command on all nodes in parallel is highly recommended, because it allows bundles to be captured in the least amount of time. For the all agents to consolidate data in the same bundle, it is important that all agents initiate capture within three minutes after the first agent initiates.

To initiate capture of service data for a specific case number, use the following syntax:

# hst capture {service} {case number} {optional: level}

The HST agent can collect data for multiple services simultaneously. To obtain the list of supported services, run the following command:

# hst list-services

Supported services:
  AMS          : Collect data for Ambari metrics issue
  Ambari       : Collect data for Ambari issue
  Falcon       : Collect data for Falcon issue
  Ganglia      : Collect data for Ganglia issue
  HBase        : Collect data for HBase issue
  HCatalog     : Collect data for HCatalog issue
  HDFS         : Collect data for HDFS issue
  Hive         : Collect data for Hive issue
  Kafka        : Collect data for Kafka issue
  Knox         : Collect data for Knox issue
  MR           : Collect data for MapReduce issue
  Nagios       : Collect data for Nagios issue
  Oozie        : Collect data for Oozie issue
  Pig          : Collect data for Pig issue
  Ranger       : Collect data for Ranger issue
  Spark        : Collect data for Spark issue
  Sqoop        : Collect data for Sqoop issue
  Storm        : Collect data for Storm issue
  Tez          : Collect data for Tez issue
  YARN         : Collect data for YARN issue
  ZK           : Collect data for ZooKeeper issue

You can specify services individually, combine services using commas as delimiters, or specify all services by using the all keyword.

Support Case Troubleshooting Capture Example

For example, to capture data just for Hadoop Distributed File System (HDFS) and for case number 0001, run hst as follows:

# hst capture HDFS 0001

To capture data for HDFS, Apaceh Hive, and Apache Oozie for case number 0002, run hst as follows:

# hst capture HDFS,HIVE,OOZIE 0002

To capture L3 capture-level data for every service listed for case number 0003, run hst as follows:

# hst capture all 0003 L3

SmartSense Analysis Capture Example

To capture data for SmartSense Analysis, only configuration and metrics are required and 0 is used as the case number:

# hst capture all 0