A place to hold mainly reading notes, and some technical stuff occasionally. 这里主要是一些读书笔记、感悟;还有部分技术相关的内容。
目录[-]
你们项目中关于上传的文件是存储在什么地方的?
我们的项目关于文件存储经过了这么一个演进过程:
FastDFS
与MinIO
以及云服务(七牛云、阿里云等的对象存储),鉴于FastDFS
配置较为复杂,最终决定使用MinIO
,易上手,可扩展。MinIO
是全球领先的对象存储先锋,在标准硬件上,读/写速度上高达183 GB / 秒 和 171 GB / 秒。MinIO用作云原生应用程序的主要存储,与传统对象存储相比,云原生应用程序需要更高的吞吐量和更低的延迟。通过添加更多集群可以扩展名称空间,更多机架,直到实现目标。同时,符合一切原生云计算的架构和构建过程,并且包含最新的云计算的全新的技术和概念。
关于对象存储,使用起来无非就是文件上传、下载与删除,再加上桶的操作而已。这里仅关注 MinIO
分布式集群的高可用性、可扩展性。
Note: 我这里使用 7.x
的版本进行实验与演示,最新版的 8.x
的 MinIO
后台管理界面有所不同,但是经过我们实际生产的测试,接口都是兼容的。
<dependency>
<groupId>io.minio</groupId>
<artifactId>minio</artifactId>
<version>7.1.4</version>
</dependency>
docker run -p 9000:9000 \
--name minio1 \
-v /opt/minio/data-single \
-e "MINIO_ACCESS_KEY=AKIAIOSFODNN7EXAMPLE" \
-e "MINIO_SECRET_KEY=wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY" \
minio/minio server /data
官方文档:https://docs.min.io/docs/deploy-minio-on-docker-compose.html
通过 docker-compose
在一台主机上,运行四个 MinIO
实例,并由 Nginx
进行反向代理,负载均衡对外统一提供服务。
涉及的两个配置:
version: '3.7'
# starts 4 docker containers running minio server instances.
# using nginx reverse proxy, load balancing, you can access
# it through port 9000.
services:
minio1:
image: minio/minio:RELEASE.2020-11-10T21-02-24Z
volumes:
- data1-1:/data1
- data1-2:/data2
expose:
- "9000"
environment:
MINIO_ACCESS_KEY: minio
MINIO_SECRET_KEY: minio123
command: server http://minio{1...4}/data{1...2}
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
minio2:
image: minio/minio:RELEASE.2020-11-10T21-02-24Z
volumes:
- data2-1:/data1
- data2-2:/data2
expose:
- "9000"
environment:
MINIO_ACCESS_KEY: minio
MINIO_SECRET_KEY: minio123
command: server http://minio{1...4}/data{1...2}
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
minio3:
image: minio/minio:RELEASE.2020-11-10T21-02-24Z
volumes:
- data3-1:/data1
- data3-2:/data2
expose:
- "9000"
environment:
MINIO_ACCESS_KEY: minio
MINIO_SECRET_KEY: minio123
command: server http://minio{1...4}/data{1...2}
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
minio4:
image: minio/minio:RELEASE.2020-11-10T21-02-24Z
volumes:
- data4-1:/data1
- data4-2:/data2
expose:
- "9000"
environment:
MINIO_ACCESS_KEY: minio
MINIO_SECRET_KEY: minio123
command: server http://minio{1...4}/data{1...2}
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
nginx:
image: nginx:1.19.2-alpine
volumes:
- ./nginx.conf:/etc/nginx/nginx.conf:ro
ports:
- "9000:9000"
depends_on:
- minio1
- minio2
- minio3
- minio4
## By default this config uses default local driver,
## For custom volumes replace with volume driver configuration.
volumes:
data1-1:
data1-2:
data2-1:
data2-2:
data3-1:
data3-2:
data4-1:
data4-2:
user nginx;
worker_processes auto;
error_log /var/log/nginx/error.log warn;
pid /var/run/nginx.pid;
events {
worker_connections 1024;
}
http {
include /etc/nginx/mime.types;
default_type application/octet-stream;
log_format main '$remote_addr - $remote_user [$time_local] "$request" '
'$status $body_bytes_sent "$http_referer" '
'"$http_user_agent" "$http_x_forwarded_for"';
access_log /var/log/nginx/access.log main;
sendfile on;
#tcp_nopush on;
keepalive_timeout 65;
#gzip on;
# include /etc/nginx/conf.d/*.conf;
upstream minio {
server minio1:9000;
server minio2:9000;
server minio3:9000;
server minio4:9000;
}
server {
listen 9000;
listen [::]:9000;
server_name localhost;
# To allow special characters in headers
ignore_invalid_headers off;
# Allow any size file to be uploaded.
# Set to a value such as 1000m; to restrict file size to a specific value
client_max_body_size 0;
# To disable buffering
proxy_buffering off;
location / {
proxy_set_header Host $http_host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_connect_timeout 300;
# Default is HTTP/1, keepalive is only enabled in HTTP/1.1
proxy_http_version 1.1;
proxy_set_header Connection "";
chunked_transfer_encoding off;
proxy_pass http://minio;
}
}
}
前台运行(可直观地查看日志):docker-compose up
后台运行:docker-compose up -d
停止服务:docker-compose down
集群运行后,可通过停止不同数量的实例观察可用性:可下载、可上传。
A stand-alone MinIO server would go down if the server hosting the disks goes offline. In contrast, a distributed MinIO setup with m servers and n disks will have your data safe as long as m/2 servers or m*n/2 or more disks are online.
For example, an 16-server distributed setup with 200 disks per node would continue serving files, up to 4 servers can be offline in default configuration i.e around 800 disks down MinIO would continue to read and write objects.
MinIO
官方建议起码要搭建一个四块盘的集群,具体配置几台机器看自己需求确定,比如:
当有4个实例的集群中4个实例都在线时,可上传、可下载。
当有4个实例的集群中3个实例都在线时,可上传、可下载(即虽然一个实例挂了,但是服务依然正常,高可用~~)。
当有4个实例的集群中仅有2个实例在线时,仅可下载,不可上传。
此时,虽然仍有1个服务实例在线,但是已经无法提供正常的上传、下载服务,即不可上传,不可下载。
我们可以得出结论:一个有4个实例运行的 MinIO
集群,有3个在线,读取写入都可以进行,有2个在线,可以保证能读取,但不能写入,若只剩下一个实例,则读写都不可进行。这与官方文档给出说明相一致(超过 m/2
个服务器在线,或者超过 m*n/2
个磁盘在线)。
Note: 在每停止一个服务后,可以查看 MinIO
后台的日志,观察各个服务实例的在线状态监测过程,方便理解。
docker exec -it 12935e3c6264 bash
这是因为 SpringBoot
项目限制了上传文件的大小,默认为1M,当用户上传了超过1M大小的文件时,就会抛出上述错误,可通过以下配置修改。
spring:
servlet:
multipart:
maxFileSize: 10MB
maxRequestSize: 30MB
docker exec -it cd34c345960c /bin/bash
无法进入容器,报错:OCI runtime exec failed: exec failed: container_linux.go:349: starting container process caused “exec: "/": permission denied”: unknown解决: docker exec -it cd34c345960c /bin/bash
改为: docker exec -it cd34c345960c sh
或者: docker exec -it cd34c345960c /bin/sh
If you have any questions or any bugs are found, please feel free to contact me.
Your comments and suggestions are welcome!