Message Image  

Customizing Docker images and Helm Charts for IIB,ACE for deployment on ICP and assigning Scaling policies for PoDs

 View Only
Mon July 06, 2020 08:33 AM

In this article we will demonstrate how to;

  • Create your own Docker image
  • Publish the Docker image to the local ICP repository
  • Load a Helm chart in an ICP catalog for the Docker image
  • Create and assign an auto-scaling policy to the deployment

Create a Docker image for ACE, IIB

Here, we illustrate the steps for an ACE image, but a similar procedure would apply to IIB as well.
  1. Download the Docker package (zip file) for ACE from Github.
  2. After unpacking the zip file, you will find the Dockerfile in the directory structure as follows:
  3. Suppose you want to customize this Docker image by including certain bar files and have them deployed when the Docker container is started. So, copy the bar files in the above directory and edit the Dockerfile as shown in the example below by adding the COPY command to copy the bars to a temporary location and define the mqsibar command to deploy the bars to the Integration Server work directory.

    # Copy in the bar file to a temporary directory
    COPY –chown=aceuser $BAR1 /tmp
    # Unzip the BAR file; need to use bash to make the profile work
    RUN bash -c ‘mqsibar -w /home/aceuser/ace-server -a /tmp/$BAR1 -c’
    # Set entrypoint to run management script
    CMD [“/bin/bash”, “-c”, “/usr/local/bin/ && IntegrationServer -w /home/aceuser/ace-server –console-log”]

  4. Now, build the Docker image.

    $ docker build -t ace:

  5. Upon successful building of the Docker image, you should see the message:

    Successfully built 514fe6a4fcc3
    Successfully tagged ace:

  6. You can verify the newly created Docker image by running it locally.

    # docker run –name myAceBar -e LICENSE=accept -P ace:
    Sourcing profile
    2018-07-23 06:40:12.773669: …..2018-07-23 06:40:13.057042: Integration server ‘ace-server’ starting initialization; version ‘’ (64-bit)
    ……………………………….2018-07-23 06:40:18.783118: About to ‘Initialize’ the deployed resource ‘Transformation_Map’ of type ‘Application’.
    2018-07-23 06:40:21.181476: About to ‘Start’ the deployed resource ‘Transformation_Map’ of type ‘Application’.
    An http endpoint was registered on port ‘7800’, path ‘/Transformation_Map’.
    2018-07-23 06:40:21.232764: The HTTP Listener has started listening on port ‘7800’ for ‘http’ connections.
    2018-07-23 06:40:21.233108: Listening on HTTP URL ‘/Transformation_Map’.
    Started native listener for HTTP input node on port 7800 for URL /Transformation_Map
    2018-07-23 06:40:22.051954: Integration server has finished initialization.
    2018-07-23 06:40:22.054202: The HTTP Listener has started listening on port ‘7600’ for ‘http’ connections.

  7. Access your Integration Server using the Admin Console. To get the port mapping information, run following command:

    We see that the admin port 7600 is mapped to 32769. So now we can access the ACE Admin Console using

    http://host IP address:32769/

  8. In order to be able to push this docker image to IBM Cloud Private (ICP), tag your image with ICP cluster information.

  9. Push the image to the ICP repository

    # docker login mycluster.icp:8500
    # docker push mycluster.icp:8500/default/ace_bar:

  10. Navigate to the ICP admin console and find the Docker image at ICP -> Catalog -> Images

Creating the Helm chart

  1. Download the base code of the Helm package for ACE from Github
  2. Unpack the downloaded zip file ( and navigate to subdirectory to get the Chart.yaml file.
  3. Edit the Chart.yaml file to give a unique name to the Helm chart that you want to create. For example,

    name: ibm-ace-bar-dev

  4. Edit the values.yaml to provide the name of the Docker image that you just pushed to the local ICP repository in step (9) in the section above. For example:

    # repository is the container repository to use, which defaults to IIB docker registry hub image
    repository: mycluster.icp:8500/default/ace_bar
    # tag is the tag to use for the container repository

  5. To verify that the Helm chart directory name/structures are correct, run the lint command as shown below. Please make sure that the name of the top level directory where the Helm package files are stored, matches with the name of the chart specified in the Chart.yaml file.

    # helm lint ibm-ace-bar-dev
    ==> Linting ibm-ace-bar-dev
    Lint OK
    1 chart(s) linted, no failures

  6. Now package the Helm chart.

    # helm package ibm-ace-bar-dev
    Successfully packaged chart and saved it to: /root/ACE/ace-helm-master/ibm-ace-bar-dev-1.0.0.tgz

  7. Load the Helm chart in the ICP catalog.

    # bx pr login -a https://mycluster.icp:8443 -u admin -p admin -c id-mycluster-account –skip-ssl-validation
    # bx pr cluster-config mycluster
    # bx pr load-helm-chart –archive ibm-ace-bar-dev-1.0.0.tgz –clustername mycluster.icp
    Loading helm chart
    Synch charts
    {“message”:”synch started”}

  8. In the ICP admin console, under the Catalog menu filter by ‘local-charts’. You should be able to see the chart that we have just published.

Creating ACE deployment using Helm charts

  1. In the ICP catalog, find the Helm package that you published as shown in the section above.
  2. Click the‘Configure’button and fill in the details. The Image repository and the image tag will be pre-filled as it comes from the values.yaml that we had updated in the section above.
  3. Click Install. The deployment process begins.
  4. Navigate to Workloads -> Helm releases. From the list, click on the Helm release that you just deployed in the step above. You would be able to see the Services, Deployment and PoD details.

  5. Click on Service name and it will provide further information like the WebUI and Integration Server Listener.
  6. The console log of your PoD can be viewed by navigating to the Deployments -> PoD -> Logs section. In the image below, we can see that the Integration Server has been started successfully.

Configuring Auto Scaling policy for Deployment

When the load on your integration server increases due to an increased volume of messages, one of the impacts you would observe is the CPU utilization and the throughput rate. In such cases you may want to scale up your integration flows horizontally to cater for the additional load so that the CPU utilization is within limits and eventually improving the message throughput rate. Also, when the peak load time is over and the message volumes are less, you would want to scale down the number of integration servers to save on CPU and memory resources. So, in a nutshell, the auto-scaling policy is required to scale-up or scale-down the number of integration servers based on certain parameters. Currently, the ICP platform provides auto-scaling based on the %CPU utilization for a given deployment.

  1. To define a scaling policy, go to Configuration -> Scaling Policies in the ICP admin console menu.
  2. Click ‘Create Policy’. It will open up a dialog box. Enter the details as shown below.

    • Provide a Name to your policy.
    • Select the Namespace that you want this to be applied to.
    • Under Scale target, provide the name of the Deployment to which you want to apply this policy. You can find the exact name of your Deployment by navigating through Workloads –> Deployments list from the left hand panel in the ICP admin console.
    • Set a value for Minimum replications which is the no. of copies of your PoD that you want to be running.
    • Set a value for Maximum replications which is the no. of copies of your PoD that you want to scale to when the CPU utilization exceeds the threshold.
    • Define the value for Target CPU which is the threshold value of % CPU Utilisation value of a Deployment above which you want the auto scaling policy to get triggered to spawn additional PoDs up to Maximum defined ones.
3. When all the values have been entered, click Create. The policy is created and gets associated with the Deployment.
4. In order to verify the working of your auto-scaling policy, you may process a large load through your Integration flows such that it causes higher CPU utilization.
5. Under Deployments section, navigate to your deployment and click the Events tab.

It would show when the replicas have scaled up and scaled down. In our example as shown below, we ran a load test for our integration flow. The Replicas scaled up from 1 to 3 when CPU utilization increased beyond the defined threshold of 10% and it scaled down to an initial value of 1 after the load test was over.