After you install Oracle Stream Analytics, it is important to setup and manage users who can use the application.

In this release of Oracle Stream Analytics, user details are stored in a database. When you create Oracle Stream Analytics schema at the time of installation, the following database tables are populated with one record in each table:

• osa_users — table containing the users

• osa_user_roles — table containing the user names and their associated roles

When you execute a query to pull in all the data from the osa_users table, you can see the following:

select * from osa_users;

+----+----------+--------------------------------------+ 
| id | username | pwd                                  | 
+----+----------+--------------------------------------+ 
|  1 | osaadmin | MD5:201f00b5ca5d65a1c118e5e32431514c | 
+----+----------+--------------------------------------+ 

where osaadmin is the pre-configured user along with the encrypted password.

When you execute a query to pull in all the data from the osa_user_roles table, you can see the following:

select * from osa_user_roles;

+---------+---------+
| user_id | role_id |
+---------+---------+
|       1 |       1 |
+---------+---------+

where role_id of value 1 indicates that the user is an administrator.

1. Click the user name in the top right corner of the screen.
2. Click System Settings.
   The System Settings page opens.
3. Click Environment.
4. Specify the server names and URLs where the Kafka Zookeeper, Yarn Resource Manager, or Spark Standalone, Path, HA Namenodes are deployed and running. Hadoop authentication is an optional setting.

   
   Description of the illustration system_settings_environment.png

   • Kafka Zookeeper Connection — the URL where the Zookeeper server or servers are configured, separated by comma. Kakfa is used as internal transport.

   • Runtime Server — the runtime server you want your Oracle Stream Analytics instance to run on

   • YARN Resource Manager URL — the URL where the YARN Resource Manager is configured, if the runtime server is Yarn

   • Spark REST URL — Spark standalone REST URL, if the runtime server is Spark standalone. To submit an Oracle Stream Analytics pipeline to Spark, the pipeline needs to be copied to a storage location that is accessible by all Spark nodes.

   • Storage — the type of storage for pipelines

   • Path — the path where the storage exists. The user who accesses this folder must have the write permission to it. The folder at this path is either created by the administrator, or Oracle Stream Analytics creates the folder at runtime.

   • HA Namenodes — If the hostname in the above URL refers to a logical HA cluster, specify the actual namenodes here in the format hostname1<port>, hostname2<port> etc.

   • Hadoop Authentication — the type of authentication configured on the Hadoop cluster. Oracle Stream Analytics supports only simple authentication.

5. Click Pipelines. Specify the various settings for the pipeline within Oracle Stream Analytics.

   • Batch Duration — the default duration of the batch for each pipeline

   • Executor Count — the default number of executors per pipeline

   • Cores per Executor — the default number of cores. A minimum value of 2 is required.

   • Executor Memory — the default allocated memory for each executor instance in megabytes

   • Cores per Driver — the default number of cores

   • Driver Memory — the default allocated memory per driver instance in megabytes

   • High Availability — toggle the default HA value as on/off for each pipeline

6. Click Proxy. If you set proper proxy, the back-end system will use these settings to reach the REST target. The proxy settings are also required for geo code related spatial patterns.
7. Click User Management. You see the list of available users here. You can add/delete users from this screen.
8. Click Save.

   Note:

   On the Yarn Cluster, make sure that you do not set a high value for yarn.nm.liveness-monitor.expiry-interval-ms (for example, 3000 ms instead of 10 minutes). This property determines the default value for how long to wait until a node manager is considered dead. Using a high value for this property does not process the events and the events are lost.