Skip to content

Latest commit

 

History

History
188 lines (149 loc) · 6.5 KB

getting-started-with-alfresco-charts.md

File metadata and controls

188 lines (149 loc) · 6.5 KB
Raw

Building your Content platform architecture with Alfresco charts

The alfresco-common library chart

Alfresco charts depend on a common library chart called alfresco-common. This chart provides a set of named templates one can use in order to build his own content platform chart.

Named templates are reusable functions that can only return strings and always take a single argument, which eventually can be an object. For example the URL helper file provides functions to extract various elements from a given URL. For example:

  • {{ template "alfresco-common.url.scheme" "https://hub.alfresco.com/" }} will return "https"
  • {{ template "alfresco-common.url.port" "https://hub.alfresco.com/" }} will return "443"

Check out the dedicated Chart documentation to get a more comprehensive list of named templates the chart offers.

Alfresco components charts

The individual Alfresco components' chart are meant to be configurable and reusable. They do not provide a working system on their own and need to be either used from an "umbrella chart" or be given additional details about other components they interact with in order to actually deliver the service they are meant to deliver. For example the alfresco-repository chart needs to have a database to initialize its content repository.

This document explains what are the options you can leverage to "pipe" an Alfresco chart with another or with an external service.

Understanding the Alfresco Content platform

Before you begin, make sure you do understand the Alfresco Content platform and its components so you know exactly what you need or do not need.

⚠️TODO: create chart diagram showing dependencies and level of dependencies

Creating your Alfresco platform

There are numerous ways to create your own Alfresco deployment and the component charts hosted in this repository can be used as building blocks for this purpose. Depending on preferences you may want to use Terraform, Kustomize or simply wrap the component charts into an "umbrella chart".

Here we'll focus on the later option but the same configuration principles apply to other methods.

In the example below we'll go through the process of integrating an ACS content platform made of:

  • Alfresco content repository
  • A database that hosted outside of the kubernetes cluster

Below example also present using an "unmbrella" Helm chart but most of the same approach is applicable to other deployment method as explained above.

Declaring chart as dependency

Helm provides a mechanism to declare dependencies and re-use existing charts.

We'll start by initializing a chart and declare the dependency to the alfresco-repository chart.

helm create myecmplatform && cd $_

Edit the Chart.yaml to give the chart a sensible description, a version and optionally specify the version of your content platform. Then declare the dependency as shown below:

dependencies:
  - name: alfresco-repository
    version: 0.1.0
    repository: https://alfresco.github.io/alfresco-helm-charts/
    alias: repo

In the example above we use repo as an alias so any value in our chart that lives under the $.repo YAML path will actually be passed to the alfresco-repository chart. In other words, values documented for the alfresco-repository chart such as replicaCount will be used from your chart as repo.replicaCount.

Configuring the chart

Now we need to let the repository where it can find its database. When it comes to configuring a chart so it can reach another component of the platform, the Alfresco component charts lets you do that in 2 different ways:

  • Using regular values
  • Using kubernetes config resources (configMap or secret)

Configuring database using values

First thing to do is to check the chart's values documentation. There, we can see the database configuration can be done using the configuration.db element in our values.yaml file as shown below:

repo:
  configuration:
    db:
      url: jdbc:postgresql//db.ecm.infra.local/alfresco
      username: ecmdaouser
      password: ecmdaopass

Configuring database using kubernetes resources

When using kubernetes resources we will want to differentiate configuration details which are sensitive and which are not. The former will be stored in configMap, while the later would use secret.

That means here we'll use a configMap to set store the database URL while credentials would live in a secret. That concretely translate into the following kubernetes resources:

apiVersion: v1
kind: ConfigMap
metadata:
  labels:
    app.kubernetes.io/instance: myEcmReleaseName
    app.kubernetes.io/name: Alfresco-content-platform
    app.kubernetes.io/version: 1.0.0
  name: ecm-database-location
  namespace: myecm
data:
  MY_ECM_DB_URL: jdbc:postgresql://db.ecm.infra.local/alfresco
  MY_ECM_DB_DRIVER: org.postgresql.Driver
apiVersion: v1
kind: Secret
metadata:
  labels:
    app.kubernetes.io/instance: myEcmReleaseName
    app.kubernetes.io/name: Alfresco-content-platform
    app.kubernetes.io/version: 1.0.0
  name: ecm-database-credentials
  namespace: myecm
data:
  MY_ECM_DB_USER: c2NvdHQ=
  MY_ECM_DB_PASS: dGlnZXI=

For example, if you're creating this confgimap using a wrapping Helm chart, you can reuse a templating function provided by the alfresco-repository chart alfresco-repository.db.cm

Now we need to let the alfresco-repository chart where to find this:

repo:
  configuration:
    db:
      existingConfigMap:
        name: ecm-database-location
      keys:
        url: MY_ECM_DB_URL
        driver: MY_ECM_DB_DRIVER
      existingSecret:
        name: ecm-database-credentials
      keys:
        username: MY_ECM_DB_USER
        password: MY_ECM_DB_PASS

If you used the templating function mentionned above and provided by the subchart the keys would be different than the ondes mentionned in the example above. Actually the function would render teh default keys so it's not needed to set them.