minio 4*4 集群 故障测试

因为一个minio有点集群故障（文件写入异常），所以基于官方的理论测试下集群容错性

一个计算规则

4*4 模式的，默认使用的纠删码条纹为16 （官方的模式是取最大，但是计算页面可以调整，对于minio来说这个是自动的）,按照套路应该是可以一个server 以及4个盘异常的，不应该出现一个节点异常造成服务不可用的
对于纠删码的奇偶校验可以自己设置，一般集群环境是4

 

 

环境准备

• docker-compose 文件

 

version: '3.7'

services:

  sidekick:

    image: minio/sidekick:v1.2.0

    tty: true

    ports:

    - "80:80"

    command: --health-path=/minio/health/ready --address :80 http://minio{1...4}:9000

  gateway:

    image:  minio/minio:RELEASE.2022-03-26T06-49-28Z

    command: gateway s3 http://sidekick --console-address ":19000"

    environment:

      MINIO_ACCESS_KEY: minio

      MINIO_SECRET_KEY: minio123

    ports:

    - "9000:9000"

    - "19000:19000"

  minio1:

    image: minio/minio:RELEASE.2022-03-26T06-49-28Z

    volumes:

      - data1-1:/data1

      - data1-2:/data2

      - data1-3:/data3

      - data1-4:/data4

    ports:

      - "9001:9000"

      - "19001:19001"

    environment:

      MINIO_ACCESS_KEY: minio

      MINIO_SECRET_KEY: minio123

    command: server http://minio{1...4}/data{1...4} --console-address ":19001"

    healthcheck:

      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]

      interval: 30s

      timeout: 20s

      retries: 3

  minio2:

    image: minio/minio:RELEASE.2022-03-26T06-49-28Z

    volumes:

      - data2-1:/data1

      - data2-2:/data2

      - data2-3:/data3

      - data2-4:/data4

    ports:

      - "9002:9000"

      - "19002:19002"

    environment:

      MINIO_ACCESS_KEY: minio

      MINIO_SECRET_KEY: minio123

    command: server http://minio{1...4}/data{1...4}  --console-address ":19002"

    healthcheck:

      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]

      interval: 30s

      timeout: 20s

      retries: 3

  minio3:

    image: minio/minio:RELEASE.2022-03-26T06-49-28Z

    volumes:

      - data3-1:/data1

      - data3-2:/data2

      - data3-3:/data3

      - data3-4:/data4

    ports:

      - "9003:9000"

      - "19003:19003"

    environment:

      MINIO_ACCESS_KEY: minio

      MINIO_SECRET_KEY: minio123

    command: server http://minio{1...4}/data{1...4} --console-address ":19003"

    healthcheck:

      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]

      interval: 30s

      timeout: 20s

      retries: 3

  minio4:

    image: minio/minio:RELEASE.2022-03-26T06-49-28Z

    volumes:

      - data4-1:/data1

      - data4-2:/data2

      - data4-3:/data3

      - data4-4:/data4

    ports:

      - "9004:9000"

      - "19004:19004"

    environment:

      MINIO_ACCESS_KEY: minio

      MINIO_SECRET_KEY: minio123

    command: server http://minio{1...4}/data{1...4} --console-address ":19004"

    healthcheck:

      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]

      interval: 30s

      timeout: 20s

      retries: 3

volumes:

  data1-1:

  data1-2:

  data1-3:

  data1-4:

  data2-1:

  data2-2:

  data2-3:

  data2-4:

  data3-1:

  data3-2:

  data3-3:

  data3-4:

  data4-1:

  data4-2:

  data4-3:

  data4-4:

故障简单测试

直接停止一个服务节点（会有4个driver 下线），比如docker-compose stop minio1 ,然后通过gateway 以及入口上传文件测试服务的读写情况，之后启动服务之后看看数据恢复情况

服务节点日志

可以查看minio4 可以看到信息

 

 

 

 

查看minio1 的文件恢复情况

 

 

说明

经过以上简单的测试，实际上minio 的集群是靠谱的，核心可能还是在nginx 入口设置有些问题，后续应该调整下，以上只是实际的问题，想验证下理论是否有问题

参考资料

https://github.com/minio/minio/blob/master/docs/distributed/SIZING.md
https://min.io/product/erasure-code-calculator?number_of_servers=8&drives_per_server=16&drive_capacity=8&stripe_size=16&parity_count=4
https://github.com/rongfengliang/minio-cluster-sidekick-learning