:sectnums:
= Batch Job on Cloud Foundry
In this demonstration, you will learn how to orchestrate short-lived data processing application (_eg: Spring Batch Jobs_) using http://cloud.spring.io/spring-cloud-task/[Spring Cloud Task] and http://cloud.spring.io/spring-cloud-dataflow/[Spring Cloud Data Flow] on Cloud Foundry.
== Using Cloud Foundry Server
=== Prerequisites
In order to get started, make sure that you have the following components:
* Local https://pivotal.io/pcf-dev[PCFDev] instance
* Local install of https://github.com/cloudfoundry/cli[cf CLI] command line tool
* Local build of https://github.com/spring-cloud/spring-cloud-dataflow[Spring Cloud Data Flow]
* Local build of Spring Cloud Data Flow's https://github.com/spring-cloud/spring-cloud-dataflow-server-cloudfoundry[Cloud Foundry Server]
* Running instance of mysql in PCFDev
NOTE: PCF 1.7.12 or greater is required to run Tasks on Spring Cloud Data Flow. As of this writing, PCFDev and PWS supports builds upon this version.
=== Running the Sample in Cloud Foundry
. Verify that CF instance is reachable
+
```
→ cf api
api endpoint: https://api.local.pcfdev.io
api version: 2.75.0
$ cf apps
Getting apps in org user-dataflow / space development as user...
OK
No apps found
```
+
. Task support needs to be enabled on pcf-dev. Being logged as `admin`, issue the following command:
+
```
cf enable-feature-flag task_creation
Setting status of task_creation as admin...
OK
Feature task_creation Enabled.
```
+
. Follow the http://docs.spring.io/spring-cloud-dataflow-server-cloudfoundry/docs/current-SNAPSHOT/reference/htmlsingle/#getting-started[getting-started] instructions to deploy Spring Cloud Data Flow's Cloud Foundry Server
+
NOTE: For this sample, all you need is the `mysql` service and in PCFDev, the `mysql` service comes with a different plan. From CF CLI, create the service by: `cf create-service p-mysql 512mb mysql` and bind this service to `dataflow-server` by: `cf bind-service dataflow-server mysql`.
+
NOTE: All the apps deployed to PCFDev start with low memory by default. It is recommended to change it to at least 768MB for `dataflow-server`. Ditto for every app spawned *by* Spring Cloud Data Flow. Change the memory by: `cf set-env dataflow-server SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_MEMORY 512`. Likewise, we would have to skip SSL validation by: `cf set-env dataflow-server SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_SKIP_SSL_VALIDATION true`.
. Once you complete the instructions, you'll be able to list the newly deployed `dataflow-server`
+
```
$ cf apps
Getting apps in org user-dataflow / space development as user...
OK
name requested state instances memory disk urls
dataflow-server started 1/1 512M 512M dataflow-server.local.pcfdev.io
```
+
NOTE: The `ORG`, `SPACE` and `USER` could vary depending on your Cloud Foundry environment in use.
. Tasks in Spring Cloud Data Flow require an RDBMS to host "task repository" (see http://docs.spring.io/spring-cloud-dataflow/docs/current-SNAPSHOT/reference/htmlsingle/#spring-cloud-dataflow-task-repository[here] for more details), so let's instruct the Spring Cloud Data Flow server to bind the `mysql` service to each deployed task:
+
```
$ cf set-env dataflow-server SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_TASK_SERVICES mysql
$ cf restage dataflow-server
```
+
NOTE: We only need `mysql` service for this sample.
. As a recap, here is what you should see as configuration for the Spring Cloud Data Flow server:
+
```
cf env dataflow-server
....
User-Provided:
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_DOMAIN: local.pcfdev.io
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_MEMORY: 512
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_ORG: pcfdev-org
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_PASSWORD: pass
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_SKIP_SSL_VALIDATION: false
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_SPACE: pcfdev-space
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_TASK_SERVICES: mysql
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_URL: https://api.local.pcfdev.io
SPRING_CLOUD_DEPLOYER_CLOUDFOUNDRY_USERNAME: user
No running env variables have been set
No staging env variables have been set
```
+
. Notice that `dataflow-server` application is started and ready for interaction via `http://dataflow-server.local.pcfdev.io` endpoint
. Connect to Spring Cloud Data Flow's `shell`.
+
```
$ cd <PATH/TO/SPRING-CLOUD-DATAFLOW>
$ java -jar spring-cloud-dataflow-shell/target/spring-cloud-dataflow-shell-<VERSION>.jar
____ ____ _ __
/ ___| _ __ _ __(_)_ __ __ _ / ___| | ___ _ _ __| |
\___ \| '_ \| '__| | '_ \ / _` | | | | |/ _ \| | | |/ _` |
___) | |_) | | | | | | | (_| | | |___| | (_) | |_| | (_| |
|____/| .__/|_| |_|_| |_|\__, | \____|_|\___/ \__,_|\__,_|
____ |_| _ __|___/ __________
| _ \ __ _| |_ __ _ | ___| | _____ __ \ \ \ \ \ \
| | | |/ _` | __/ _` | | |_ | |/ _ \ \ /\ / / \ \ \ \ \ \
| |_| | (_| | || (_| | | _| | | (_) \ V V / / / / / / /
|____/ \__,_|\__\__,_| |_| |_|\___/ \_/\_/ /_/_/_/_/_/
<VERSION>
Welcome to the Spring Cloud Data Flow shell. For assistance hit TAB or type "help".
server-unknown:>
```
+
. Connect `shell` with the `server` running at `http://dataflow-server.local.pcfdev.io`
+
```
server-unknown:>dataflow config server http://dataflow-server.local.pcfdev.io
Successfully targeted http://dataflow-server.local.pcfdev.io
```
+
. Let's build and register the batch-job https://github.com/spring-cloud/spring-cloud-task/tree/master/spring-cloud-task-samples/batch-job[example] from Spring Cloud Task samples. For convenience, the final https://github.com/spring-cloud/spring-cloud-dataflow-samples/raw/master/tasks/simple-batch-job/batch-job-1.0.0.BUILD-SNAPSHOT.jar[uber-jar artifact] is provided with this sample
+
```
dataflow:>app register --type task --name simple_batch_job --uri https://github.com/spring-cloud/spring-cloud-dataflow-samples/raw/master/tasks/simple-batch-job/batch-job-1.3.0.BUILD-SNAPSHOT.jar
```
+
. Create the task with `simple-batch-job` application
+
```
dataflow:>task create foo --definition "simple_batch_job"
```
NOTE: Unlike Streams, the Task definitions don't require explicit deployment. They can be launched on-demand, scheduled, or triggered by streams.
+
. Verify there's *still* no Task applications running on PCFDev - they are listed only after the initial launch/staging attempt on PCF
+
```
cf apps
Getting apps in org pcfdev-org / space pcfdev-space as user...
OK
name requested state instances memory disk urls
dataflow-server started 1/1 768M 512M dataflow-server.local.pcfdev.io
```
+
. Let's launch `foo`
+
```
dataflow:>task launch foo
```
+
. Verify the execution of `foo` by tailing the logs
+
```
→ cf logs foo
Retrieving logs for app foo in org pcfdev-org / space pcfdev-space as user...
2016-08-14T18:48:54.22-0700 [APP/TASK/foo/0]OUT Creating container
2016-08-14T18:48:55.47-0700 [APP/TASK/foo/0]OUT
2016-08-14T18:49:06.59-0700 [APP/TASK/foo/0]OUT 2016-08-15 01:49:06.598 INFO 14 --- [ main] o.s.b.c.l.support.SimpleJobLauncher : Job: [SimpleJob: [name=job1]] launched with the following parameters: [{}]
...
...
2016-08-14T18:49:06.78-0700 [APP/TASK/foo/0]OUT 2016-08-15 01:49:06.785 INFO 14 --- [ main] o.s.b.c.l.support.SimpleJobLauncher : Job: [SimpleJob: [name=job1]] completed with the following parameters: [{}] and the following status: [COMPLETED]
...
...
2016-08-14T18:49:07.36-0700 [APP/TASK/foo/0]OUT 2016-08-15 01:49:07.363 INFO 14 --- [ main] o.s.b.c.l.support.SimpleJobLauncher : Job: [SimpleJob: [name=job2]] launched with the following parameters: [{}]
...
...
2016-08-14T18:49:07.53-0700 [APP/TASK/foo/0]OUT 2016-08-15 01:49:07.536 INFO 14 --- [ main] o.s.b.c.l.support.SimpleJobLauncher : Job: [SimpleJob: [name=job2]] completed with the following parameters: [{}] and the following status: [COMPLETED]
...
...
2016-08-14T18:49:07.71-0700 [APP/TASK/foo/0]OUT Exit status 0
2016-08-14T18:49:07.78-0700 [APP/TASK/foo/0]OUT Destroying container
2016-08-14T18:49:08.47-0700 [APP/TASK/foo/0]OUT Successfully destroyed container
```
NOTE: Verify `job1` and `job2` operations embeddded in `simple-batch-job` application are launched independently and they returned with the status `COMPLETED`.
+
NOTE: Unlike LRPs in Cloud Foundry, tasks are short-lived, so the logs aren't always available. They are generated only when the Task application runs; at the end of Task operation, the container that ran the Task application is destroyed to free-up resources.
+
. List Tasks in Cloud Foundry
+
```
→ cf apps
Getting apps in org pcfdev-org / space pcfdev-space as user...
OK
name requested state instances memory disk urls
dataflow-server started 1/1 768M 512M dataflow-server.local.pcfdev.io
foo stopped 0/1 1G 1G
```
+
. Verify Task execution details
+
```
dataflow:>task execution list
╔══════════════════════════╤══╤════════════════════════════╤════════════════════════════╤═════════╗
║ Task Name │ID│ Start Time │ End Time │Exit Code║
╠══════════════════════════╪══╪════════════════════════════╪════════════════════════════╪═════════╣
║foo │1 │Sun Aug 14 18:49:05 PDT 2016│Sun Aug 14 18:49:07 PDT 2016│0 ║
╚══════════════════════════╧══╧════════════════════════════╧════════════════════════════╧═════════╝
```
. Verify Job execution details
+
```
dataflow:>job execution list
╔═══╤═══════╤═════════╤════════════════════════════╤═════════════════════╤══════════════════╗
║ID │Task ID│Job Name │ Start Time │Step Execution Count │Definition Status ║
╠═══╪═══════╪═════════╪════════════════════════════╪═════════════════════╪══════════════════╣
║2 │1 │job2 │Sun Aug 14 18:49:07 PDT 2016│1 │Destroyed ║
║1 │1 │job1 │Sun Aug 14 18:49:06 PDT 2016│1 │Destroyed ║
╚═══╧═══════╧═════════╧════════════════════════════╧═════════════════════╧══════════════════╝
```
+
== Summary
In this sample, you have learned:
* How to register and orchestrate Spring Batch jobs in Spring Cloud Data Flow
* How to use the `cf` CLI in the context of Task applications orchestrated by Spring Cloud Data Flow
* How to verify task executions and task repository
