Table of Contents

Driven User Guide

version 1.1.1

Configuring Teams for Collaboration

Applications are not deployed and monitored in isolation. Usually, application development is a collaboration between the IT developers, engineering, business analysts, and other data scientists. When applications become operational, it becomes vital to create collaboration to investigate job behavior and observed characteristics.

In addition, in any reasonably-sized cluster, it becomes vital to categorize applications for various needs such as governance, compliance, developing models for charge-backs, etc. For example, in a bank you may have the Hadoop production cluster as a single enterprise capability, shared between different lines of businesses such as marketing, fraud, operational risk, and across organizations in different geographies. In such case, you may want to monitor applications along such categories to enforce compliance, track usage, develop chargeback models, compare historical behavior across the same class of application. The comparison can be done over variables that are not only available and aggregated through Job dashboard, but also through application-defined counters. Finally, visualizing applications within the same class makes it easier to identify outliers (such as certain jobs taking larger time).

Creating Your Team

The Teams feature in Driven allows you to create groups – or Teams in our parlance - and associate applications to these teams. Before you create a new team, click on the Team icon to display all existing teams. You can then assess the your applications and apply the appropriate team structure.

Team_Name Figure 1: Create a Team by entering a name.

Perform the following steps to create a team:

Step 1: Create a Team

Perform the following tasks to create a team:

  1. In the top right menu, click the admin icon to display the pop-up menu.

  2. Select Team list. The My Teams window appears.

  3. In the Create a new Team field, enter your team name.

  4. Click Create Team. In the My Teams pane, your new team name appears in the list.

  5. In the My Teams list, click the link of your new team name. The details of your team appears. You can now add members to your team.

Associated with your new team name is an API key. All applications executed with that particular key will show up under the Team. More details on the API key in Step 4.

Step 2: Create Users

After creating the team, you should have only one existing member name associated with the new team, the admin user. Before you can add new members to your team you need to create a user.

Perform the following tasks to create a user:

  1. Go to the admin menu>Teams list><your team name>.

  2. Click Create a new user. The Create User window appears.

  3. Enter the user’s email, name, password, and password confirmation. In the Team field, use the drop-down menu to select the team name you want to associate the user.

  4. Click the checkbox to enable the user to be an administrator for this team.

  5. Click Click User.

Create-User

Figure 2: Create a user who is a member on your team.

Step 3: Add Users to the Team

In Step 2, you have created a user and added the user to your team. In this step, you are adding the newly created user to other teams if appropriate.

Perform the following tasks to add a user to the team:

  1. Go to the admin menu>Team list> All Team list.

  2. Click the link of the desired team name.

  3. In the Invite a user field, enter the user’s email address.

  4. Click Invite User.

Note
 A notification appears if the SMTP service is not enabled in your Driven server. Your invitation will need to resent once the SMTP service is enabled.

Invite_User

Figure 3. Add the user to your Team.

Step 4: Run the application

As previously mentioned in Step 1, before you can run the application you need to set an API key that is associated with your Team when an Hadoop application is executed. Based on the specific level (cluster, user, or application) that your Hadoop application is based, set the API key by using the following:

Cluster level - You can create cascading-service.properties in your Hadoop configuration directory (referred later as $HADOOP_CONF). For Hadoop 1.x, this may be $HADOOP_HOME/conf and for Hadoop 2.x, it will be $HADOOP_INSTALL/etc/hadoop

$ echo cascading.management.document.service.apikey=_YOUR_API_KEY_ >> $HADOOP_CONF/cascading-service.properties

User level - You can set the environment variable ‘DRIVEN_API_KEY’

$ export DRIVEN_API_KEY=_YOUR_API_KEY_

Application level - You can set the Java JVM level system property named, cascading.management.document.service.apikey:

$ java -Dcascading.management.document.service.apikey=_YOUR_API_KEY_ ....

Step 5: View your application in Teams

Apps_2_Teams

Figure 4. Select applications associated with your Team.

Your Team is now categorized, which not only improves collaboration, but better organization in a multi-tenant cluster environment.

Note
 As the admin user, you will see both My Team and All Teams. However, if you are a user with no admin permissions, then you will only see My Team.

Next

Using Search for deep-insights