125: OpenIM 集群化部署提案 [RFC] Cluster Deployment for OpenIMSDK

RFC #0007 Cluster Deployment for OpenIMSDK

Meta

• Name: Cluster Deployment for OpenIMSDK
• Start Date: 2023-09-04
• Author(s): @cubxxw
• Status: Draft
• RFC Pull Request: https://github.com/OpenIMSDK/community/edit/main/RFC/0007-openim-cluster-deployment.md
• OpenIMSDK Pull Request: (leave blank)
• OpenIMSDK Issue: https://github.com/OpenIMSDK/Open-IM-Server/issues/1002
• Supersedes: N/A

📇Topics

• RFC #0000 Cluster Deployment for OpenIMSDK
  • Meta
  • Summary
  • Definitions
  • Motivation
  • What it is
  • How it Works
  • Migration
  • Drawbacks
  • Alternatives
  • Prior Art
  • Unresolved Questions
  • Spec. Changes (OPTIONAL)
  • History

Summary

This proposal aims to provide a streamlined approach for deploying OpenIMSDK in a cluster environment, considering multiple deployment strategies, including but not limited to helm, sealos, and k3s.

Definitions

• Cluster: A set of computers that work together to run a set of applications.
• Kubernetes (K8s): An open-source system for automating deployment, scaling, and management of containerized applications.
• Helm: A package manager for Kubernetes.
• Sealos: An efficient Kubernetes installer.

Motivation

• To streamline and standardize the OpenIMSDK deployment in various cluster environments.
• To leverage the scalability, redundancy, and high availability offered by cluster deployments.
• To provide clear documentation and guidance for deploying OpenIMSDK in a cluster.

What it is

This feature aims to provide a guideline and tools required to deploy OpenIMSDK in a cluster. Target personas include OpenIMSDK developers, platform operators, and users who wish to deploy OpenIMSDK in a scalable and efficient manner.

How it Works

1. K8s Environment Setup: Choose the desired cluster setup tool, such as k3s, sealos, or helm. Set up the cluster according to the respective tool's documentation.
2. Service Deployment: Deploy OpenIMSDK server components followed by chat components using the respective Kubernetes configurations.
3. Configuration & Environment Variables: Adjust configuration files and environment variables based on the deployment environment and requirements.
4. Scaling & Load Balancing: Based on demand, scale the services either manually or automatically using Kubernetes' horizontal pod autoscaler.
5. Monitoring & Logging: Use tools like Prometheus and Grafana to monitor the health and performance of the services.

Migration

Users currently running OpenIMSDK in non-clustered environments can migrate their data to the new cluster environment. Detailed steps will be provided, including backup, transfer, and restore procedures.

Drawbacks

• Increased complexity in setup and maintenance.
• Requires expertise in Kubernetes and the chosen cluster deployment tools.

Alternatives

• Continue to support traditional non-cluster deployments.
• Explore other cloud-native deployment tools beyond the mentioned in this proposal.

Prior Art

Several contributors and even the official team have previously provided cluster deployment solutions:

• k8s-jenkins
• Open-IM-Server-k8s-deploy
• openim-charts
• deploy-openim

Unresolved Questions

• Which deployment strategy is the most efficient and reliable for OpenIMSDK?
• How do we ensure a seamless migration experience for users transitioning from a non-clustered to a clustered environment?

Spec. Changes (OPTIONAL)

There may be a need to introduce new configurations or environment variables specific to clustered deployments. This section will be updated as these changes are identified during the implementation phase.

History

• https://github.com/OpenIMSDK/Open-IM-Server/issues/601

Fix Issue:

• https://github.com/OpenIMSDK/Open-IM-Server/issues/688
• https://github.com/OpenIMSDK/Open-IM-Server/issues/621
• https://github.com/OpenIMSDK/Open-IM-Server/issues/459