MinioServer启动模式
MinioServer启动模式
1. 简介
minio支持多种server启动模式:
standalone模式
即要管理的磁盘都在host本地。该启动模式一般仅用于实验环境、测试环境的验证和学习使用。在standalone模式下,还可以分为
- non-erasure code mode
- erasure code mode。
1.1 standalone 之 non-erasure code 模式
所谓non-erasure code mode,即minio server启动时仅传入一个本地磁盘目录参数:比如:
$minio server data
Endpoint: http://10.10.126.88:9000 http://127.0.0.1:9000
AccessKey: minioadmin
SecretKey: minioadmin
Browser Access:
http://10.10.126.88:9000 http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://10.10.126.88:9000 minioadmin minioadmin
... ...
在这样的启动模式下,对于每一份对象数据,minio直接在data下面存储这份数据,不会建立副本,也不会启用纠删码机制。因此,这种模式无论是服务实例还是磁盘都是“单点”,无任何高可用保障,磁盘损坏就表示数据丢失。
1.2 standalone 之 non-erasure code 模式
erasure code mode即为minio server实例传入多个本地磁盘参数。一旦遇到多于一个磁盘参数,minio server会自动启用erasure code mode。erasure code对磁盘的个数是有要求的(最少4个),如不满足要求,实例启动将失败:
$minio server data1 data2
ERROR Invalid command line arguments: Incorrect number of endpoints provided [data1 data2]
> Please provide an even number of endpoints greater or equal to 4
HINT:
For more information, please refer to https://docs.min.io/docs/minio-erasure-code-quickstart-guide
erasure code启用后,要求传给minio server的endpoint(standalone模式下,即本地磁盘上的目录)至少为4个。minio server启用纠删码机制后,会自动将传入的disk drive划分为多个erasure coding set,每个erasure coding set中的disk drive的数量可以是:4, 6, 8, 10, 12, 14 和16。minio server会根据传入disk drive的数量自动计算set个数和每个set中的disk drive数量。比如下面例子中,我们传入四个endpoint(disk drive)给minio server:
$minio server data1 data2 data3 data4
Formatting 1 zone, 1 set(s), 4 drives per set.
WARNING: Host local has more than 2 drives of set. A host failure will result in data becoming unavailable.
Status: 4 Online, 0 Offline.
Endpoint: http://10.10.126.88:9000 http://127.0.0.1:9000
AccessKey: minioadmin
SecretKey: minioadmin
Browser Access:
http://10.10.126.88:9000 http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://10.10.126.88:9000 minioadmin minioadmin
... ...
从minio server的输出日志来看,minio server将这些drive放入了一个erasure coding set了。在输出日志中,我们还看到一行WARNING: Host local has more than 2 drives of set. A host failure will result in data becoming unavailable.,即minio server警告我们:这个erasure coding set中有多于两个的drive都在local host上,这样一旦host宕机,那么数据将无法获取。(每个set 有4个drive,根据纠删码的机制,这个set的最大允许失效的disk数量为4/2=2)。
1.2.1 语法糖-省略号
$minio server data{1...18}
Formatting 1 zone, 3 set(s), 6 drives per set.
WARNING: Host local has more than 3 drives of set. A host failure will result in data becoming unavailable.
WARNING: Host local has more than 3 drives of set. A host failure will result in data becoming unavailable.
WARNING: Host local has more than 3 drives of set. A host failure will result in data becoming unavailable.
Status: 18 Online, 0 Offline.
Endpoint: http://10.10.126.88:9000 http://127.0.0.1:9000
AccessKey: minioadmin
SecretKey: minioadmin
Browser Access:
http://10.10.126.88:9000 http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://10.10.126.88:9000 minioadmin minioadmin
... ...
minio server data{1...18}
等价于
minio server data1 data2 data3 data4 data5 data6 data7 data8 data9 data10 data11 data 12 data13 data14 data15 data16 data17 data18
minio server会自行扩展省略号代表的内容。我们看到:当我们传入18个disk drive后,minio server创建了3个erasure coding set,每个set中有6个disk drive。同样,minio server还针对每个set输出了一行WARNING:每个Set中有三个以上的disk drive都位于同一台host上。
这些WARNING我们可以通过distributed mode来解决。顾名思义,distributed mode下,minio server实例和其管理的disk drive分布在多台host上,这种模式可以避免minio server实例单点,数据也将分布在不同host上的不同disk中,实现了高可用,提升了整体的容灾能力。由于处理多个host上的disk,distribute mode默认就会启动erasure coding set机制。
1.3 distributed mode
在distributed mode下,minio server后面的远程的endpoint采用http url编码格式:
export MINIO_ACCESS_KEY=<ACCESS_KEY>
export MINIO_SECRET_KEY=<SECRET_KEY>
$minio server http://host{1...4}:9000/minio/data{1...4}
上面例子中的minio server命令相当于4个host,每个host上启动一个minio server实例,每个实例都管理16的disk drive(包括本地和远程的)。上述命令等价于:
$minio server http://host1:9000/minio/data1 http://host1:9000/minio/data2 http://host1:9000/minio/data3 http://host1:9000/minio/data4 http://host2:9000/minio/data1 http://host2:9000/minio/data2 http://host2:9000/minio/data3 http://host2:9000/minio/data4 http://host3:9000/minio/data1 http://host3:9000/minio/data2 http://host3:9000/minio/data3 http://host3:9000/minio/data4 http://host4:9000/minio/data1 http://host4:9000/minio/data2 http://host4:9000/minio/data3 http://host4:9000/minio/data4
minio同样会自动将这些disk drive划分为若干个erasure coding set。每个endpoint用http://address/disk-drive-path
的形式编码。注意:这条命令在host1、host2、host3和host4上都要执行
zone
的概念
2. 比如下面这个例子:
$minio server data{1...8} data{9...16}
Formatting 1 zone, 1 set(s), 8 drives per set.
WARNING: Host local has more than 4 drives of set. A host failure will result in data becoming unavailable.
Formatting 2 zone, 1 set(s), 8 drives per set.
WARNING: Host local has more than 4 drives of set. A host failure will result in data becoming unavailable.
Status: 16 Online, 0 Offline.
Endpoint: http://10.10.126.88:9000 http://127.0.0.1:9000
AccessKey: minioadmin
SecretKey: minioadmin
Browser Access:
http://10.10.126.88:9000 http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://10.10.126.88:9000 minioadmin minioadmin
... ...
我们在命令行中给minio server传入两组采用“省略号”语法的参数,minio认为每组就是一个**“zone”,这里有两组,因此minio创建了两个zone**。在每个zone内,minio创建了一个erasure coding set,每个set中有8个disk drive。对于外部的写数据请求,minio server会首先查找可用空间多的zone,然后再在zone内选择set和disk drive。
如果不用“省略号”语法,那么minio server会将后面传入的所有disk drive放入一个zone中
3. 部署distributed minio集群
3. 1. 单机上部署distributed minio集群
我们的验证环境采用最小的distributed minio模式:单机、one zone, one erasure coding set, 4 disk drive。下面是部署的示意图:
我们没有使用“省略号”语法,在单机上不是很好模拟。我们通过下面脚本来启动该minio集群:
# cat startup_minio.sh
#!/bin/bash
export MINIO_ACCESS_KEY="minio"
export MINIO_SECRET_KEY="minio123"
for i in {01..04}; do
nohup minio server --address ":90${i}" http://127.0.0.1:9001/root/minio-install/data1 http://127.0.0.1:9002/root/minio-install/data2 http://127.0.0.1:9003/root/minio-install/data3 http://127.0.0.1:9004/root/minio-install/data4 > "/root/minio-install/90${i}.log"& 2>&1
done
启动该minio集群,并查看启动状态:
# bash startup_minio.sh
# ps -ef|grep minio
root 1218 1 11 21:58 pts/5 00:00:01 minio server --address :9001 http://127.0.0.1:9001/root/minio-install/data1 http://127.0.0.1:9002/root/minio-install/data2 http://127.0.0.1:9003/root/minio-install/data3 http://127.0.0.1:9004/root/minio-install/data4
root 1219 1 11 21:58 pts/5 00:00:01 minio server --address :9002 http://127.0.0.1:9001/root/minio-install/data1 http://127.0.0.1:9002/root/minio-install/data2 http://127.0.0.1:9003/root/minio-install/data3 http://127.0.0.1:9004/root/minio-install/data4
root 1220 1 3 21:58 pts/5 00:00:00 minio server --address :9003 http://127.0.0.1:9001/root/minio-install/data1 http://127.0.0.1:9002/root/minio-install/data2 http://127.0.0.1:9003/root/minio-install/data3 http://127.0.0.1:9004/root/minio-install/data4
root 1221 1 11 21:58 pts/5 00:00:01 minio server --address :9004 http://127.0.0.1:9001/root/minio-install/data1 http://127.0.0.1:9002/root/minio-install/data2 http://127.0.0.1:9003/root/minio-install/data3 http://127.0.0.1:9004/root/minio-install/data4
root@instance-cspzrq3u:~/minio-install# ls
9001.log 9002.log 9003.log 9004.log data1 data2 data3 data4 startup_minio.sh
root@instance-cspzrq3u:~/minio-install# tail -100f 9001.log
Formatting 1 zone, 1 set(s), 4 drives per set.
Attempting encryption of all config, IAM users and policies on MinIO backend
Status: 4 Online, 0 Offline.
Endpoint: http://192.168.16.4:9001 http://172.17.0.1:9001 http://172.18.0.1:9001 http://127.0.0.1:9001
Browser Access:
http://192.168.16.4:9001 http://172.17.0.1:9001 http://172.18.0.1:9001 http://127.0.0.1:9001
.... ...
3.2 mc配置与管理
minio官方提供了mc命令行工具,用于对minio server进行管理。我们首先要为mc创建一个管理本地minio server(:9001)的配置:
# mc config host add myminio
这里我们使用mc添加了一个所谓”host”,指向上面创建的minio server(:9001)。上面的命令实质上是在~/.mc/config.json中写入了如下配置:
# cat ~/.mc/config.json
{
"version": "9",
"hosts": {
"myminio": {
"url": "http://localhost:9001",
"accessKey": "minio",
"secretKey": "minio123",
"api": "s3v4",
"lookup": "auto"
}
}
}
接下来,我们通过mc命令在minio集群中添加三个bucket:
root@instance-cspzrq3u:~# mc mb myminio/image
Bucket created successfully `myminio/image`.
root@instance-cspzrq3u:~# mc mb myminio/video
Bucket created successfully `myminio/video`.
root@instance-cspzrq3u:~# mc mb myminio/audio
Bucket created successfully `myminio/audio`.
root@instance-cspzrq3u:~# mc ls myminio
[2020-03-16 15:19:55 CST] 0B audio/
[2020-03-16 15:19:48 CST] 0B image/
[2020-03-16 15:19:52 CST] 0B video/
新创建的bucket默认的访问policy是none,即外部无访问权限:
root@instance-cspzrq3u:~# mc policy get myminio/image
Access permission for `myminio/image` is `none`
根据我们的设计,我们需要给这三个bucket添加外部可读取权限,以image这个bucket为例:
root@instance-cspzrq3u:~# mc policy set download myminio/image
Access permission for `myminio/image` is set to `download`
root@instance-cspzrq3u:~# mc policy get myminio/image
Access permission for `myminio/image` is `download`
3.3. load balancer设置
这里我们使用一个nginx前置在minio集群外部,下面是为minio创建的nginx配置文件(/etc/nginx/conf.d/minio.conf):
// /etc/nginx/conf.d/minio.conf
upstream minio_cluster {
server localhost:9001;
server localhost:9002;
server localhost:9003;
server localhost:9004;
}
server {
listen 9000;
server_name myminio.tonybai.com;
# 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 X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header Host $http_host;
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_cluster;
}
location /image/ {
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_set_header Host $http_host;
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;
client_max_body_size 1000m;
proxy_buffering off;
proxy_pass http://minio_cluster;
}
}
重启nginx(nginx -s reload)。
我们使用浏览器访问一下http://myminio.tonybai.com:9000/
,登录后,你将看到如下页面:
选择左侧的”image” bucket,点击右下角的”+”号,我们可以上传一张图片:gopher-daily-logo.png,上传后,我们退出登录。然后通过地址http://myminio.tonybai.com:9000/image/gopher-daily-logo.png
访问该图片。你也可以通过wget命令下载该图片:
$wget -c http://myminio.tonybai.com:9000/image/gopher-daily-logo.png
--2020-03-16 15:40:20-- http://myminio.tonybai.com:9000/image/gopher-daily-logo.png
正在解析主机 myminio.tonybai.com (myminio.tonybai.com)... 106.12.69.83
正在连接 myminio.tonybai.com (myminio.tonybai.com)|106.12.69.83|:9000... 已连接。
已发出 HTTP 请求,正在等待回应... 200 OK
长度:59736 (58K) [image/png]
正在保存至: “gopher-daily-logo.png”
gopher-daily-logo.png 100%[============================================>] 58.34K 253KB/s 用时 0.2s
2020-03-16 15:40:20 (253 KB/s) - 已保存 “gopher-daily-logo.png” [59736/59736])
4 开启多版本控制
要开启多版本控制,必须开启纠删码模式
开启纠删码模式后,我们上传的文件可以看到历史版本
我们再存储可以看到他的多版本文件