2060 Override Cluster and Machine Pipelines¶

20 Minutes intermediate

Create application specific pipeline overrides based on standardized pipelines

In this lab, the User will create a Pipeline and use it to build clusters. Business ROI: Using standardized pipelines helps ensure consistent operational results without restricting business imperatives for use-case specific needs.

Prerequisites¶

Additional Labs:

1020

Concepts¶

clusters
pipelines
params

Summary¶

Add the Dev Library Content Pack¶

Navigate to the Catalog
Find the dev-library using the search bar
Install or Update dev-library if not loaded, by clicking the install button

Create a Pipelines¶

Navigate to Pipelines

Click the Create button The form should have the following:

* Set `Name` to `lab2060`
* Set the Param `universal-application` to `lab2060`
  * The value must be the part after `universal-application-` in the `Name` field.
* Set the Param `universal-application-chain-index-override` to `runbook`
  * This will set the pipeline to apply a configurable application as the last step of the pipeline.
* Set the `icon` and `color` fields to your choice.
* Click `Save`

Click Add Params
Select dev/wait-icons by typing
Select universal/runbook-during-flexiflow by typing

Edit the dev/wait-icons parameter Make the list of values:

```json
  [
    "cat",
    "arrow up",
    "arrow right",
    "arrow down",
    "arrow left",
    "dog"
  ]
```

Click the little `save` icon

Edit the universal/runbook-during-flexiflow parameter Make the values:

```json
  [ "wait-time" ]
```

Click the little `save` icon

This injects the `wait-time` task into the runbook workflow element of the pipeline.

Look at the Pipeline Object¶

Review the Runbook Workflow Entrypoints This view shows the Workflows chains in the Pipeline.

Containers and VMs will enter through `universal-start` Workflow.

Bare Metal and PXEing VMs will use the `universal-discover` Workflow.

Review the JSON This view shows the raw JSON object for the Pipeline. The view can also show YAML.
```
A button in the upper right, `Enable Editing`, allows for direct editing of the object.
```

Create a Cluster to validate the Pipeline¶

Navigate to Clusters

Click Add Set Name to lab2060

Set the `broker/name` to `context-broker`

Set the `broker/set-pipeline` to `universal-application-lab2060`

Click `Save`

Review the Machines Created by the Cluster¶

Navigate to Machines

Notice the Machines with names starting with lab2060- The machines were created by the context-broker for the lab2060 clusters.

These machines were assigned the `universal-application-lab2060` pipeline and are running
that now.  The icons of the machine should be cycling through the list in the pipeline.

The task is set to `wait-time`.

Clean up the cluster¶

Navigate to Clusters
Select the lab2060 cluster
Click Actions and select Workflow Mode Clusters must be Workflow Modee to be removed.
Click Actions and select Delete (Cleanup)
Click Cleanup to confirm The cluster will run a workflow to cleanup the machines by asking the broker to remove them. Once complete, the cluster will be removed.

Create Cluster Pipeline¶

Navigate to Pipelines

Click the Create button The form should have the following:

* Set `Name` to `lab2060-cluster`
* Set the Param `universal-application` to `lab2060-cluster`
  * The value must be the part after `universal-application-` in the `Name` field.
* Set the Param `universal-application-chain-index-override` to `cluster`
  * This will set the pipeline to follow set of workflows to create a cluster.
* Set the Meta required params field to `broker/name`
  * This will require that the cluster page ask for that field.  This is a space-separated list.
* Click `Save`

Click Add Params
Select broker/set-pipeline by typing
Select universal/runbook-during-flexiflow by typing
Edit the broker/set-pipeline parameter The value should be set to universal-application-lab2060.
```
Click the little `save` icon
```

Edit the universal/runbook-during-flexiflow parameter Make the values:

```json
  [ "wait-time" ]
```

Click the little `save` icon

This injects the `wait-time` task into the runbook workflow element of the pipeline.

Create a Cluster to validate the Cluster Pipeline¶

Navigate to Clusters

Click Add Set Name to lab2060

Set `Cluster Pipeline` to `lab2060-cluster`

**NOTICE** The required parameters changed.

Set the `broker/name` to `context-broker`

Do NOT set the `broker/set-pipeline` parameter

Click `Save`

Navigate to the lab2060 Activity view Notice that this time the cluster continues on to running the wait-time task after creating the cluster machines.
Navigate to Machines If this is done quickly, the machines will not have necessarily completed all there initialization and pipeline completion. The Cluster pipeline defaults to not waiting for machines to complete their pipelines before the cluster continues. This can be changed.

Clean up the cluster¶

Navigate to Clusters
Select the lab2060 cluster
Click Actions and select Workflow Mode Clusters must be Workflow Modee to be removed.
Click Actions and select Delete (Cleanup)
Click Cleanup to confirm The cluster will run a workflow to cleanup the machines by asking the broker to remove them. Once complete, the cluster will be removed.

Update Cluster Pipeline to Wait for Machines¶

Navigate to lab2060-cluster Pipeline
Click Add Params
Select universal/cluster-provision-post-flexiflow by typing

Edit the universal/cluster-provision-post-flexiflow parameter Make the values:

```json
  [ "cluster-wait-for-members" ]
```

Click the little `save` icon

This injects the `cluster-wait-for-members` task into the cluster provision workflow element of the pipeline.

This shows injection of tasks.  Alternatively, there is a parameter for this already, `cluster/wait-for-members`.
Setting this to `true` instead of this will cause the same effect.

Create a Cluster to validate the Cluster Pipeline¶

Navigate to Clusters

Click Add Set Name to lab2060

Set `Cluster Pipeline` to `lab2060-cluster`

**NOTICE** The required parameters changed.

Set the `broker/name` to `context-broker`

Do NOT set the `broker/set-pipeline` parameter

Click `Save`

Navigate to the lab2060 Activity view Notice that this time the cluster runs cluster-wait-for-members and pauses waiting for the machines to finish.
Navigate to Machines If this is done quickly, the machines will not have necessarily completed all there initialization and pipeline completion. Once all the machines finish, the cluster will then run the wait-time task.

Clean up the cluster¶

Navigate to Clusters
Select the lab2060 cluster
Click Actions and select Workflow Mode Clusters must be Workflow Modee to be removed.
Click Actions and select Delete (Cleanup)
Click Cleanup to confirm The cluster will run a workflow to cleanup the machines by asking the broker to remove them. Once complete, the cluster will be removed.