Skip to main content

​
Install AutoMQ on OCI

This document will guide you through the process of deploying AutoMQ for Kafka on Oracle Cloud Infrastructure (OCI). We will cover environment preparation and installation steps to ensure a smooth deployment.

​
1. Prepare the VPC Network

First, we need a correctly configured Virtual Cloud Network (VCN) to host the Kubernetes cluster and AutoMQ instances. Please ensure your VCN meets the following requirements:
  • Same Region: The VCN must be in the same OCI region where you plan to deploy the OKE cluster and your business applications.
  • Subnets:
    • The VCN must contain at least two subnets, ideally one Public Subnet and at least one Private Subnet.
    • Public Subnet: Used for deploying resources that require public internet access, such as Load Balancers.
    • Private Subnet: Used for deploying OKE worker nodes and AutoMQ instances to enhance security.
    • Ensure each subnet has sufficient available IP addresses to meet the scaling needs of your OKE cluster and applications.
    • Reference: Managing VCNs and Subnets
  • Internet Gateway: Attach an Internet Gateway to your Public Subnet and configure appropriate routing rules to allow traffic from the internet.
  • NAT Gateway: Attach a NAT Gateway to your Private Subnet and configure appropriate routing rules. This allows resources within the Private Subnet (e.g., OKE nodes) to access the internet for downloading images and dependencies, without allowing direct internet access to these resources.
  • Service Gateway: Enable a Service Gateway for your VCN and configure routing rules to allow resources within the private network to access OCI Object Storage via OCI’s backbone network. This ensures secure and high-performance data transfer.

​
2. Prepare Object Storage

AutoMQ innovatively offloads streaming data to cloud provider object storage for disaggregated storage and compute, achieving extreme cost-efficiency. In an OCI environment, each AutoMQ cluster requires two separate object storage buckets:
  • Data Bucket: Used for storing Kafka message data. Recommended naming convention: automq-data-<cluster-name>.
  • Ops Bucket: Used for storing operational data such as cluster logs and metrics. Recommended naming convention: automq-ops-<cluster-name>.
Please create these two buckets in the same OCI region as your OKE cluster.

​
3. Prepare the Kubernetes Cluster (OKE)

AutoMQ is a cloud-native messaging and streaming system designed to run entirely on Kubernetes. Therefore, you need to prepare an Oracle Kubernetes Engine (OKE) cluster.

​
3.1. Create OKE Cluster

When creating your OKE cluster, pay attention to the following key parameter configurations:
  • Kubernetes Version: It is recommended to choose a recent stable version, for example, v1.33.x. Please refer to the official OCI documentation for the latest supported versions.
  • Network Mode:
    • Select “VCN-Native Pod Networking” (NPN) network mode.
    • In the network configuration, select the VCN you prepared in the first step and configure the Kubernetes API Endpoint and Worker Nodes subnets to the correct private subnets.
    • Configure Network Security Groups (NSGs) to allow necessary traffic into and out of the cluster.
  • Reference: Creating a Kubernetes Cluster

​
3.2. Create Node Pool

For optimal performance and resource isolation, it is recommended to create a dedicated node pool for AutoMQ.
  • Instance Type (Shape): We recommend VM.Standard.E5.Flex shapes, which offer flexible CPU and memory ratios.
  • Resource Configuration:
    • OCPUs: 2
    • Memory: 16 GB
  • Node Count (Size): Select at least 3 nodes to ensure high availability.

​
4. Deploy AutoMQ with Helm Chart

Once all the above preparatory work is complete, you can proceed to deploy AutoMQ Enterprise using Helm Chart. For detailed installation steps, parameter configurations (such as choosing between S3WAL and EBSWAL modes), and advanced configurations, please refer to our standard Helm Chart deployment documentation. Deploy AutoMQ Enterprise with Helm Chart▸