Product Description

Oracle SQL Connector for HDFS is a part of an Oracle Big Data Connectors Software. It can transfers data from a Hadoop Distributed File System (HDFS) to an Oracle Database.

Known Versions

• 2.3
• 3.0
• 3.1

Software Pattern Summary

Platforms Supported by the Pattern

The pattern supports only UNIX platforms.

Identification

Software Instance Triggers

Simple Identification Mappings

The following components are identified using simple identity mappings.

Obtaining key variables

Obtaining Hadoop Install Root

Hadoop install root is obtained from the trigger process command line using the following regular expression:

• 
  ^(/.+)/bin/hadoop

Obtaining Oracle HDFS Install Root

Oracle HDFS install root is obtained from the trigger process arguments using the following regular expression:

• 
  (?:^|\s)(/\S+)/jlib/orahdfs.jar

Versioning

Version information for the product is currently collected using one of two possible methods.

The methods are tried in an order of precedence based on likely success and/or accuracy of the information that can be gathered. Once a result is obtained, the methods lower in precedence are not attempted. In order of precedence the methods are:

Path Versioning

BMC Atrium Discovery extracts version information from the Oracle HDFS install root using the following regular expression:

• /orahdfs-(\d+(?:\.\d+)+)

File Versioning

If the version information wasn't obtained using previous methods, BMC Atrium Discovery extracts it from the README.txt file located in the <Oracle HDFS install root>/doc/ directory  using the following regular expression:

• Release\s+(\d+(?:\.\d+)+)\s+README file

Application Model Produced by Software Pattern

Product Architecture

The Oracle SQL Connector for HDFS is installed and configured on top of the system with the Orace database and, if necessary, on the Hadoop client system that is used to access Hive tables. The HDFSConnector models relationship with the Oracle database and Hadoop SIs.

Software Pattern Model

The Oracle SQL Connector for HDFS pattern triggers on the hadoop process with the arguments that contain the following substring:

• oracle.hadoop.extttab.HdfsStream

SI Depth

The pattern produces an instance based Software Instance for Oracle SQL Connector for HDFS, with the SI key that is created using the SI type (Oracle SQL Connector for HDFS) and the key of the host that runs the Oracle SQL Connector for HDFS instance.

Software Instance Modeling

The model of the Oracle SQL Connector for HDFS deployment generated with this pattern:

Relationship with Oracle database

The pattern searches for the Oracle Database Server Software Instance installed on the same host as Oracle SQL Connector for HDFS. If only one instance is found, it creates a communication relationship between these two SIs. If more then one is identified, the pattern uses an SQL query to retirev the OSCH_BIN_PATH value. If the value corresponds to the Oracle SQL Connector for HDFS install root, the pattern creates a communication relationship between the Oracle SQL Connector for HDFS SI and the Oracle Database Server SIs.

Relationship with Hadoop

If Hadoop install root is retrieved, the pattern searches for the Apache Hadoop DataNode Software Instance with matching install root on the same host as Oracle SQL Connector for HDFS. If this SI is found, the pattern creates a communication relationship between these two SIs.

Subject Matter Expertise

Subject Matter Expert input will be welcome on any other potential approaches not discussed above to improving product versioning coverage and depth of Oracle SQL Connector for HDFS and components.

Testing

The pattern for Oracle SQL Connector for HDFS was tested using actual installations of BMC Atrium Discovery running on Linux.

Information Sources

• Oracle® Big Data Connectors User's Guide Release 4

Open Issues

N/A

Created By: Chris Blake (10 December 2014)