首页 🙇‍♂️ DevOps

GlusterFS服务总结一下就是启动守护进程一个、卷组进程两个,挂载进程一个;端口分别是守护进程端口24007,卷组进程端口49152,后续端口+1.

1 GlusterFS守护进程

/usr/sbin/glusterd -p /var/run/glusterd.pid --log-level INFO

端口:24007,111

2 启动一个zzy卷组后,进程为

/usr/sbin/glusterfsd -s glusterfs-nodes-01 --volfile-id zzy.glusterfs-nodes-01.glusterfs-zzy -p /var/run/gluster/vols/zzy/glusterfs-nodes-01-glusterfs-zzy.pid -S /var/run/gluster/31c02248ffc86608.socket --brick-name /glusterfs/zzy -l /var/log/glusterfs/bricks/glusterfs-zzy.log --xlator-option *-posix.glusterd-uuid=f2295e80-8b9a-4cc1-bdc2-1ae5c89fdafc --process-name brick --brick-port 49152 --xlator-option zzy-server.listen-port=49152

端口:49152

/usr/sbin/glusterfs -s localhost --volfile-id shd/zzy -p /var/run/gluster/shd/zzy/zzy-shd.pid -l /var/log/glusterfs/glustershd.log -S /var/run/gluster/78e6d8ba530b2cec.socket --xlator-option *replicate*.node-uuid=f2295e80-8b9a-4cc1-bdc2-1ae5c89fdafc --process-name glustershd --client-pid=-6

这是一个监控进程!

3 再启动一个uusafe卷组后,进程为

/usr/sbin/glusterfsd -s glusterfs-nodes-01 --volfile-id uusafe.glusterfs-nodes-01.glusterfs-uusafe -p /var/run/gluster/vols/uusafe/glusterfs-nodes-01-glusterfs-uusafe.pid -S /var/run/gluster/04672df42c2d0842.socket --brick-name /glusterfs/uusafe -l /var/log/glusterfs/bricks/glusterfs-uusafe.log --xlator-option *-posix.glusterd-uuid=f2295e80-8b9a-4cc1-bdc2-1ae5c89fdafc --process-name brick --brick-port 49153 --xlator-option uusafe-server.listen-port=49153

端口:49153

4 挂载后启动的进程为

/usr/sbin/glusterfs --process-name fuse --volfile-server=glusterfs-nodes-01 --volfile-id=zzy /zzy
/usr/sbin/glusterfs --process-name fuse --volfile-server=glusterfs-nodes-01 --volfile-id=uusafe /uusafe

5 状态信息

gluster volume status all
Status of volume: uusafe
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick glusterfs-nodes-01:/glusterfs/uusafe  49153     0          Y       3738
Brick glusterfs-nodes-02:/glusterfs/uusafe  49153     0          Y       3495
Self-heal Daemon on localhost               N/A       N/A        Y       3706
Self-heal Daemon on glusterfs-nodes-02      N/A       N/A        Y       3469

Task Status of Volume uusafe
------------------------------------------------------------------------------
There are no active volume tasks

Status of volume: zzy
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick glusterfs-nodes-01:/glusterfs/zzy     49152     0          Y       3685
Brick glusterfs-nodes-02:/glusterfs/zzy     49152     0          Y       3448
Self-heal Daemon on localhost               N/A       N/A        Y       3706
Self-heal Daemon on glusterfs-nodes-02      N/A       N/A        Y       3469

Task Status of Volume zzy
------------------------------------------------------------------------------
There are no active volume tasks
gluster volume info

Volume Name: uusafe
Type: Replicate
Volume ID: 77b5eb26-1be9-4edd-ace2-8505f8b6051d
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: glusterfs-nodes-01:/glusterfs/uusafe
Brick2: glusterfs-nodes-02:/glusterfs/uusafe
Options Reconfigured:
performance.client-io-threads: off
nfs.disable: on
storage.fips-mode-rchecksum: on
transport.address-family: inet

Volume Name: zzy
Type: Replicate
Volume ID: 1c85cc79-9aec-4267-90be-bd6b12900d4c
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: glusterfs-nodes-01:/glusterfs/zzy
Brick2: glusterfs-nodes-02:/glusterfs/zzy
Options Reconfigured:
performance.client-io-threads: off
nfs.disable: on
storage.fips-mode-rchecksum: on
transport.address-family: inet

6 堆栈结构

  • gluster:是cli命令执行工具,主要功能是解析命令行参数,然后把命令发送给glusterd模块执行。
  • glusterd:是一个管理模块,处理gluster发过来的命令,处理集群管理、存储池管理、brick管理、负载均衡、快照管理等。集群信息、存储池信息和快照信息等都是以配置文件的形式存放在服务器中,当客户端挂载存储时,glusterd会把存储池的配置文件发送给客户端。
  • glusterfsd:是服务端模块,存储池中的每个brick都会启动一个glusterfsd进程。此模块主要是处理客户端的读写请求,从关联的brick所在磁盘中读写数据,然后返回给客户端。
  • glusterfs:是客户端模块,负责通过mount挂载集群中某台服务器的存储池,以目录的形式呈现给用户。当用户从此目录读写数据时,客户端根据从glusterd模块获取的存储池的配置文件信息,通过DHT算法计算文件所在服务器的brick位置,然后通过Infiniband RDMA 或Tcp/Ip 方式把数据发送给brick,等brick处理完,给用户返回结果。存储池的副本、条带、hash、EC等逻辑都在客户端处理。




文章评论

目录