Replacing OCP Nodes Using OceanBase Admin Toolkit (OAT): A Step‑by‑Step Guide

OceanBase Cloud Platform (OCP) is an enterprise‑grade database management platform built on OceanBase.

In production environments OCP is created first and then used to manage and monitor clusters, so installing OCP is usually the first step of a system rollout; later hardware changes may require node replacement.

This article describes two methods for replacing OCP nodes; the first uses the OceanBase Admin Toolkit (OAT) and the second uses the antman script. The following sections focus on the OAT method.

Note: The author's OCP load balancer uses F5, so the new machines must be configured in F5 first; other load‑balancer scenarios are analogous.

Environment Background

If you have experience with an OB production environment you may know that early OCP versions required three Docker packages—OCP software, metadb, and OBProxy—while later versions combine DB and Proxy into a single Docker image. OAT can only manage the combined DB+Proxy package; the separate metadb still requires the antman script for replacement.

Software Information

Software

Version

OCP

ocp-all-in-one:3.3.3-20220906114643

metadb+proxy

OB2277_OBP320_x86_20220429

OAT

4.1.1_20230519_x86

Operation Process

3.1 Environment Check / Preparation

Check the replacement machine environment, including disk partitioning, creating an admin user, and installing Docker. After installation verify the setup.

cd /root/t-oceanbase-antman/clonescripts/
sh precheck.sh -m ocp

Install OAT Tool

# 1. Upload and extract OAT package
tar -xvf oat-all-in-one-x86-411.tar

# 2. Load OAT image
csocpth:~# cd oat-all-in-one-x86/
csocpth:~/oat-all-in-one-x86# docker load -i oat_4.1.1_20230519_x86.tgz
... (output omitted) ...

# 3. Run OAT container
oat_image=`docker images | grep oat | head -1 | awk '{printf $1":"$2"