...
Table of Contents |
---|
Topology Components
Jenkins Host: Where the Jenkins controller (providing build/deploy services as part of ALM) is run
...
Other Information Server Environment(s): Downstream Information Server environments, including testing and (optionally) Production. These environments can be MettleCI deployment targets without requiring the deployment of any MettleCI components.
MettleCI Agent Host: A MettleCI-dedicated Windows server hosting an IBM DataStage Client tier which is used by your Build system's agent, in conjunction with the MettleCI Command Line Interface, to automate build and deployment activities.
Topology Connections
The following high-level MettleCI architecture shows the key software components to be installed, and their communications between hosts.
Gliffy | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
The above diagram is one possible configuration, and is presented as an example only.
...
#
...
Summary
...
Description
...
Technical Notes
...
1
...
MettleCI Workbench Service (Engine Tier) ↔︎ Your Work item Management system
...
The MettleCI Workbench application running on your DataStage Engine tier needs to perform a dynamic lookup of
...
Work items when displaying the Git Commit
...
The Workbench service on your DataStage Engine tier needs to be able to use your Work Item Management system’s query API, over whatever protocols/ports it requires.
...
2
...
page.
The MettleCI Workbench application running on your DataStage Engine tier needs to
...
7
...
commit to your Git platform.
...
3
...
Jenkins Host ↔︎ Your Git repository
...
Jenkins can keep pipeline definitions (as Jenkinsfile assets) in a repository so they can be versioned the same as the artefacts.
...
4
...
Jenkins Host ↔︎ Developer Workstation DataStage designer client
...
Web access to Jenkins by developers. Similar web access should exist to the work item management and git repository
...
5
...
Jenkins Host ↔︎ MettleCi Host DataStage designer client
...
Jenkins controller issues commands to the Jenkins Agent to perform MettleCI related tasks such as code movement, and DataStage deployment (project manipulation, job compilation)
...
6
...
Developer Workstation ↔︎ IS/DS Service/Engine tiers
...
Normal developer activity such as editing jobs, test compilations, etc
The Jenkins Host and your Git repository need to communicate in order to configure Git commits to automatically trigger Jenkins pipelines. Jenkins then needs access to your Git repository to source the artefacts for build and deployment. The Jenkins pipeline definitions ('Jenkinsfiles') are themselves stored and managed in a Git repository.
The Developer Workstation provides data engineers with access to your ALM tools' user interfaces via a supported web browser.
The Developer Workstation provides data engineers with access to your Jenkins user interfaces via a supported web browser.
The Developer Workstation provides a DataStage client tier with access to the development environment's DataStage Engine and Services tiers.
Your Jenkins pipeline perform their duties via an agent installed on the MettleCI Host. (Read about Jenkins ports and services here)
The MettleCI Host requires regular access to the development environment's DataStage Engine
...
8
...
Developer Workstation DataStage designer client ↔︎ MettleCI Host
...
The (optional) plugins effect checkin, test case specification generation, and <?> via the Mettle CI host, further the developer can access the (optional) Azkaban scheduler and wallboard components via a supported web browser
...
9
...
and Services tiers, identical to that of Developer Workstations.
The MettleCI Host requires regular access to the downstream test environments' DataStage Engine and Services tiers, to
...
affect automated deployment.
Using Multiple Engines and Multiple Jenkins Agents
In real-world scenarios it’s possible that security restrictions or network zoning prohibit a single MettleCI Agent Host from communicating with DataStage environments across development, testing, and production domains. In this case you may need to consider a deployment topology using multiple Jenkins Agents, each running on a dedicated MettleCI Agent Host.
Gliffy | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
The diagram above (explained in more detail on this page) describes how you might deploy Jenkins Agents in complex multi-environment DataStage topologies.
This is by no means the only way to support multiple DataStage environments, but is certainly one of the easiest solutions to understand and maintain.