Running distributed

Run the "NumberedCat" application in a distributed environment.

In this tutorial, you will:

Submit the "NumberedCat" application from the Stream processing tutorial
Verify the health of the application PEs
Verify the application output
Cancel the job from the distributed Streams instance.

Note: When the "NumberedCat" application was compiled, the SPL compiler not only created a stand-alone executable file, but also a Streams application bundle (sab) which can be submitted to a distributed Streams instance.

Before you begin

Complete the Stream processing tutorial.

Obtain a started Streams domain and instance:

If using the QSE, validate that the instance is up and running.

If not using the QSE, use these commands to create, configure, and start a one node domain and instance:

streamtool mkdomain --embeddedzk      # make a domain
streamtool startdomain --embeddedzk   # start the domain
streamtool genkey                     # generate keys 
streamtool mkinstance --embeddedzk    # make an instance
streamtool startinstance --embeddedzk # start the runtime instance

Procedure

Set up your environment for Streams.
In your command terminal, source the streamsprofile.sh file under Streams installation directory. For example:
```
source streams-install-directory/bin/streamsprofile.sh
```
Navigate to numberedcat directory:
```
cd numberedcat
```
Remove the old result file:
```
rm -f data/result.txt
```
Submit the application to the Streams instance and make note of the job ID.
For example, use the following command:
```
streamtool submitjob output/NumberedCat.sab -P file="cat.txt"
```
Note: For this and future streamtool commands, if you need to specify a domain name other than the default, add -d domain-name. If you need to specify an instance name other than the default, add -i instance-name.

The command outputs several messages including what the job ID is.
Verify application PE health.
For example, use the following command:
```
streamtool lspes -j job-id
```
The command outputs the list of PEs for the job, their health, what resource it's running on, and more. The PE(s) should all be Healthy.

Important: If PEs are not healthy, inspect the PE logs using streamtool viewlog --print --pe pe-id.
Verify output:
```
cat data/result.txt
```
Similar to the stand-alone run, the distributed program creates a file called result.txt in the numberedcat/data directory containing the numbered lines of cat.txt.
Cancel the job:
```
streamtool canceljob job-id
```
The command tells the Streams instance to stop and clean up all PEs and the job. Jobs run indefinitely until cancelled.

What to do next

If you created a Streams and instance for this tutorial and no longer need it, clean up the environment:

streamtool stopinstance --embeddedzk # stop the runtime instance
streamtool rminstance --embeddedzk   # remove the runtime instance
streamtool stopdomain --embeddedzk   # stop the domain
streamtool rmdomain --embeddedzk     # remove the domain

Continue to the Types and functions tutorial to learn how to create user-defined SPL types and functions to improve code reuse.