README.md

Utils to initialize Hadoop core Java objects

For now it only supports HDFS and so documentation is only limited to this case. Will be extended as new features are added.

Usage

In your spring configuration class add annotation

  import org.trustedanalytics.utils.hdfs.EnableHdfs;

  @Configuration
  @EnableHdfs
  public class SomeConfigClass {
  }

Now you can use in some class:

  ...
  
  @Autowired
  private HdfsConfig hdfsConfig;

  public void someMethod() {
    FileSystem fs = hdfsConfig.getFileSystem();
    ...
  }
  
  ...

Which FileSystem is created depends on Spring profile that is active (to set profile, set SPRING_PROFILES_ACTIVE, for example: export SPRING_PROFILES_ACTIVE=local)

cloud

This is profile that is usually activated in Cloud Foundry. It reads:

• credentials for authentication/identity from user provided service ("ups") bound to application (i.e. kerberos-service). If "ups" does not exists in your space, you can create it with commands:

In environment with kerberos

  cf cups kerberos-service -p '{ "kdc": "kdc-host", "kpassword": "kerberos-password", "krealm": "kerberos-realm", "kuser": "kerberos-user" }'
  cf bs <app_name> kerberos-service

In environment with simple authentication method set

  cf cups identity-service -p '{ "kuser": "user-name" }'
  cf bs <app_name> identity-service

• other parameters that are provided by binding to HDFS service instance

local

It works on a local file system. If the env variable FOLDER is set than this is the one that is going to be used. Otherwise, temporary folder will be created that will be removed on normal application termination.