Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 4 Next »

The preparations described below are required to install  on Amazon Web Services using Helm charts and Docker images. Read through the steps below and follow each step before installing. This pre-installation is normally not needed to be done again, once you have done it.

Prerequisites

The following must be installed before you can install:


ApplicationDownload fromComments
Helm version 3.x
https://helm.sh/docs/intro/installMandatory
Cert-manager 1.1.0 or laterhttps://cert-manager.io

Recommended

Note, if you do not want to use cert-manager, there is an option to do the Helm install with the parameter mzOperator.webhook.tls.cert.delegate=internal . Search for documented helm values to understand the implications of this.

aws clihttps://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.htmlMandatory, for configuring your eks cluster with kubeconfig
kubectlhttps://kubernetes.io/docs/tasks/tools/install-kubectl/Mandatory, for all administrative tasks with the cluster
Terraform v12https://learn.hashicorp.com/terraform/getting-started/install.htmlOptional, if you want to spin up the new infrastructure and use the example templates in the upcoming sections
eksctlhttps://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.htmlOptional, if you want to spin up the new infrastructure and use the example templates in the upcoming sections

is installed with Derby as the default platform database. Should you want to run your installation on PostgreSQL, you must install the following:

  • AWS RDS with an engine PostgreSQL. or Use your existing database instance with right connection and credentials. That should be able to create the database structure for .

The existing Infrastructure should have the following services before you install.

  • EKS Cluster or Opeshift Cluster
  • Worker Nodes
  • VPC (Public and Private nodes)
  • NAT Gateways
  • Internet Gateways
  • Routing tables 
  • EFS as a persistent storage

    Optional services, in case they are needed:
  • Load Balancers (Application Load Balancers)
  • RDS Instance preferably postgres

Deployment Architecture 

VPC - Virtual Private Cloud

Two or three availablity zones (AZ) are supported for the Virtual Private Cloud, which is the recommendation. Each AZ has one public and one private subnet. Note that the number of nodes does not relate to the number of AZs. You can run two nodes with a VPC in three AZs. The nodes only occupy two AZs, but have the ability to switch to another AZ if one goes down. Using three AZs is better for high availability and redundancy. In terms of cost, each AZ adds one NAT Gateway (if nodes are in private subnets). The recommended configuration is to have all nodes in the private subnets. This is more secure and does not allow direct access to the machine as they do not have any public IP assigned.

Kubernetes

This is an overview of the different pods and load balancers. A standard setup of load balancers in public subnets and kubernetes pods in private subnets are used.

If you do not have any existing setup or starting with fresh account in AWS, Please refer the Setup (3.0) and modify accordingly to match with your production environment.

Initialization

  1. Use the aws configure command to setup your AWS credentials.

    $ aws configure
    
    aws_access_key_id = <access key id>
    aws_secret_access_key = <secret access key>
    output = json
    region = <region where you have your EKS cluster>
    


    [hide] Line Brake [/hide]

  2. Update kubeconfig context to access your eks cluster.

    $ aws eks update-kubeconfig --name <name_of_your_cluster>
    
    To set the default cluster, used as the example in this installation:
    $ aws eks update-kubeconfig --name mz-eks

    Note!

    This step can be omitted if you do not have an existing eks cluster or if you will be using the Openshift Cluster.

    If you do not have an eks cluster installed, you can create one in the Setup (3.0).


    [hide] Line Brake [/hide]

  3. verify that your are in right cluster and can list the nodes and other resources

    $ kubectl get node
    $ kubectl get all

    Note!

    This step can be omitted if you do not have an existing eks cluster or if you will be using the Openshift Cluster.

    If you do not have an eks cluster installed, you can create one in the Setup (3.0).



    [hide] Line Brake [/hide]

  4. To verify that the Helm CLI is initialized, run the following command:

    $ helm version -c

    Output example:

    $ helm version -c
    version.BuildInfo{Version:“v3.2.1”, GitCommit:“fe51cd1e31e6a202cba7dead9552a6d418ded79a”, GitTreeState:“clean”, GoVersion:“go1.13.10"}

This section contains:

  • No labels