Image Layering Best Practices

Prev Next

search by keyword

Recent Pages

Documentation

Download PDF

Image Layering Best Practices

Last updated: 2023-05-19 16:05:27

Download PDF

Overview

This document describes how to build and manage business images in layers and provides best practices for managing container images of all types using TCR.

Advantages of container image layering

Resources are shared to improve the utilization.
Image management is standardized to facilitate DevOps implementation.
TCR's Ops-free image acceleration easily makes large-scale image distribution faster by 5-10 times.
TCR Enterprise has been accessed to Tencent CloudAudit. You can check the logs of read and write operations of instances, namespaces, and image repositories in "Event History".

Prerequisites

Before using a private image managed in TCR for application deployment, you need to complete the following preparations:

You have created a TCR Enterprise instance in the TCR console. If you haven't done so, create one first. For more information, see Creating an Enterprise Edition Instance.
If you are using a sub-account, you must have granted the sub-account operation permissions for the corresponding instance. For more information, see Example of Authorization Solution of TCR Enterprise.

 Note: This also applies to the existing TCR instances. You only need to modify the image repository address.

1. F3S Docker Files Overview

The project consists of the following parts:

$ tree -L 3 ./f3s-docker-files
./f3s-docker-files
├── README.md                                       ------  README file
├── DockerBuildImages.sh                            ------  Image build script
├── 0.base                                          ------  0. Build various types of system images at the base layer
│   ├── alpine                                      ------    Build the alpine system image at the base layer
│   │   └── Dockerfile
│   └── centos-7.8                                  ------    Build the CentOS 7.8 system image at the base layer
│       ├── Dockerfile
│       └── centos-7.8.2003-x86_64-docker.tar.xz
├─  1.ops                                           ------  1. Build various types of images at the Ops layer
│   └── Dockerfile-alpine                           ------    Build the alpine image at the Ops layer
├── 2.lang                                          ------  2. Build various types of images at the language layer
│   └── Dockerfile-alpine-kona                      ------    alpine-kona image at the language layer 
└── 3.app                                           ------  3. Build various types of images at the application layer
    ├── jmeter
    │   ├── Dockerfile-jmeter-base                  ------    Build the jmeter-base image
    │   ├── Dockerfile-jmeter-grafana-reporter      ------    Build the jmeter-grafana-reporter image
    │   ├── Dockerfile-jmeter-master                ------    Build the jmeter-master image
    │   └── Dockerfile-jmeter-slave                 ------    Build the jmeter-slave image
    ├── nginx
    │   ├── Dockerfile-alpine-nginx                 ------    Build the alpine-nginx image
    │   ├── default.conf
    │   └── nginx.conf
    └── skywalking
        └── Dockerfile-alpine-kona-skywalking       ------    Build the alpine-kona-skywalking image

alpine/Dockerfile: Build with the official Alpine 3.13 Docker image to support common Ops tools.
centos-7.8/Dockerfile: Build with the official CentOS 7.8 Docker image to support common Ops tools.
Dockerfile-alpine-kona: Build with the Dockerfile-alpine and TencentKona 8.0.5 binary package. The Kona is partially trimmed to control the image size.
Dockerfile-jmeter-base: Build based on the official JMeter 5.4.1 binary package.
Dockerfile-jmeter-grafana-reporter: Build based on Grafana-Reporter to generate JMeter PDF reports from Grafana dashboards.
Dockerfile-jmeter-master: Build based on Jmeter-base to implement distributed master stress test with JMeter.
Dockerfile-jmeter-slave: Build based on Jmeter-base to implement distributed slave stress test with JMeter.
Dockerfile-alpine-nginx: Build with Dockerfile-alpine to add NGINX configuration initialization and logging specifications.
Dockerfile-alpine-kona-skywalking: Build with the official Dockerfile-alpine-kona and SkyWalking 8.5 binary package.