GlusterFS高可用集群部署及K8S对接实战篇
作者：尹正杰
版权声明：原创作品，谢绝转载！否则将追究法律责任。
目录
一.GlusterFS存储解决方案概述
1.glusterfs概述
Gluster是一种软件定义的分布式存储，可以扩展到数PB。它为对象、块和文件存储提供接口。
官网地址:
https://www.gluster.org/
github地址:
https://github.com/gluster/glusterfs
glusterFS架构图:
https://docs.gluster.org/en/latest/Quick-Start-Guide/Architecture/
2.GlusterFS环境准备
| 主机名 | 操作系统 | IP地址 | 磁盘 |
|---|---|---|---|
| master241 | Ubuntu 22.04 LTS | 10.0.0.241 | sdb ---> 300G（磁盘容量可以更大） |
| worker242 | Ubuntu 22.04 LTS | 10.0.0.242 | sdb ---> 300G |
| worker243 | Ubuntu 22.04 LTS | 10.0.0.243 | sdb ---> 300G |
如上表所示，我采用的是Ubuntu 22.04LTS系统，3个节点来部署GlusterFS集群。
3.Ubuntu安装GlusterFS
温馨提示:
官方对于Ubuntu仅支持如下版本:
Ubuntu 16.04 LTS、18.04 LTS、20.04 LTS、20.10、21.04
尽管官方这样说，添加PPA可能会失败，但并不影响安装Glusterfs。
参考链接:
https://docs.gluster.org/en/latest/Install-Guide/Install/#for-ubuntu
https://docs.gluster.org/en/latest/Install-Guide/Community-Packages/#community-packages
实操案例:
1.安装通用的软件包
apt install software-properties-common
2.添加社区GlusterFS PPA(如果Ubuntu 22.04LTS报错404可以忽略，继续执行下一步)
add-apt-repository ppa:gluster/glusterfs-7
apt update
3.安装glusterfs
apt install glusterfs-server
4.配置Glusterfs开机自启动
systemctl enable --now glusterd.service
5.检查GlusterFS是否正常启动
systemctl is-enabled glusterd.service
systemctl status glusterd.service
4.GlusterFS集群配置
1.查看集群状态
[root@master241 ~]# gluster peer status
Number of Peers: 0
[root@master241 ~]#
2.将其他节点加入到现有的集群环境
[root@master241 ~]# gluster peer probe 10.0.0.242
peer probe: success
[root@master241 ~]#
[root@master241 ~]# gluster peer probe 10.0.0.243
peer probe: success
[root@master241 ~]#
3.集群节点查看集群状态
[root@master241 ~]# gluster peer status
Number of Peers: 2
Hostname: 10.0.0.242
Uuid: a2922996-e4cc-4331-a78b-fb3dbf7234e4
State: Peer in Cluster (Connected)
Hostname: 10.0.0.243
Uuid: 301a57f8-21a0-49d0-808f-35b99a061541
State: Peer in Cluster (Connected)
[root@master241 ~]#
[root@worker242 ~]# gluster peer status
Number of Peers: 2
Hostname: 10.0.0.241
Uuid: dc3b2a67-388d-485b-a1e3-302dc0cc8687
State: Peer in Cluster (Connected)
Hostname: 10.0.0.243
Uuid: 301a57f8-21a0-49d0-808f-35b99a061541
State: Peer in Cluster (Connected)
[root@worker242 ~]#
[root@worker243 ~]# gluster peer status
Number of Peers: 2
Hostname: 10.0.0.241
Uuid: dc3b2a67-388d-485b-a1e3-302dc0cc8687
State: Peer in Cluster (Connected)
Hostname: 10.0.0.242
Uuid: a2922996-e4cc-4331-a78b-fb3dbf7234e4
State: Peer in Cluster (Connected)
[root@worker243 ~]#
[root@worker243 ~]#
参考链接：
https://docs.gluster.org/en/latest/Install-Guide/Configure/
二.GlusterFS基本管理
1.各节点准备磁盘
1.各节点创建挂在点
[root@master241 ~]# mkdir -pv /data/glusterfs/
[root@worker242 ~]# mkdir -pv /data/glusterfs/
[root@worker243 ~]# mkdir -pv /data/glusterfs/
2.格式化磁盘
[root@master241 ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
...
sdb 8:16 0 300G 0 disk
sdc 8:32 0 500G 0 disk
sdd 8:48 0 1T 0 disk
sr0 11:0 1 1024M 0 rom
[root@master241 ~]#
[root@master241 ~]# mkfs.xfs -i size=512 /dev/sdb
[root@worker242 ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
...
sdb 8:16 0 300G 0 disk
sdc 8:32 0 500G 0 disk
sdd 8:48 0 1T 0 disk
sr0 11:0 1 1024M 0 rom
[root@worker242 ~]#
[root@worker242 ~]# mkfs.xfs -i size=512 /dev/sdb
[root@worker243 ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
...
sdb 8:16 0 300G 0 disk
sdc 8:32 0 500G 0 disk
sdd 8:48 0 1T 0 disk
sr0 11:0 1 1024M 0 rom
[root@worker243 ~]#
[root@worker243 ~]# mkfs.xfs -i size=512 /dev/sdb
3.配置开机挂载数据盘
[root@master241 ~]# echo "/dev/sdb /data/glusterfs xfs defaults 0 0" >> /etc/fstab
[root@worker242 ~]# echo "/dev/sdb /data/glusterfs xfs defaults 0 0" >> /etc/fstab
[root@worker243 ~]# echo "/dev/sdb /data/glusterfs xfs defaults 0 0" >> /etc/fstab
4.刷新fstab并验证
[root@master241 ~]# mount -a && df -h | grep data
/dev/sdb 300G 2.2G 298G 1% /data/glusterfs
[root@master241 ~]#
[root@worker242 ~]# mount -a && df -h | grep data
/dev/sdb 300G 2.2G 298G 1% /data/glusterfs
[root@worker242 ~]#
[root@worker243 ~]# mount -a && df -h | grep data
/dev/sdb 300G 2.2G 298G 1% /data/glusterfs
[root@worker243 ~]#
温馨提示:
建议大家在各节点准备单独的磁盘进行挂载，便于后续创建复制卷，如果不这样做的话，后续的命令在执行时必须加"--force"。
官方的建议是单独搞一块磁盘单独存储GlusterFS数据，不建议直接使用根分区进行存储。
参考链接:
https://docs.gluster.org/en/latest/Install-Guide/Configure/#partition-the-disk
2.创建复制卷
1.创建复制卷
[root@master241 ~]# gluster volume create yinzhengjie-volume replica 3 10.0.0.241:/data/glusterfs/r1 10.0.0.242:/data/glusterfs/r2 10.0.0.243:/data/glusterfs/r3
volume create: yinzhengjie-volume: success: please start the volume to access data
[root@master241 ~]#
2.查看复制卷状态
[root@master241 ~]# gluster volume status yinzhengjie-volume
Volume yinzhengjie-volume is not started
[root@master241 ~]#
3.启动复制卷
[root@master241 ~]# gluster volume start yinzhengjie-volume
volume start: yinzhengjie-volume: success
[root@master241 ~]#
4.再次查看复制卷状态
[root@master241 ~]# gluster volume status yinzhengjie-volume
Status of volume: yinzhengjie-volume
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick 10.0.0.241:/data/glusterfs/r1 56017 0 Y 16318
Brick 10.0.0.242:/data/glusterfs/r2 54370 0 Y 10750
Brick 10.0.0.243:/data/glusterfs/r3 51212 0 Y 10368
Self-heal Daemon on localhost N/A N/A Y 16335
Self-heal Daemon on 10.0.0.242 N/A N/A Y 10767
Self-heal Daemon on 10.0.0.243 N/A N/A Y 10385
Task Status of Volume yinzhengjie-volume
------------------------------------------------------------------------------
There are no active volume tasks
[root@master241 ~]#
5.查看卷的详细信息
[root@master241 ~]# gluster volume info
Volume Name: yinzhengjie-volume
Type: Replicate
Volume ID: f962bd43-e4d4-424b-b693-52714afb6aed
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x 3 = 3
Transport-type: tcp
Bricks:
Brick1: 10.0.0.241:/data/glusterfs/r1
Brick2: 10.0.0.242:/data/glusterfs/r2
Brick3: 10.0.0.243:/data/glusterfs/r3
Options Reconfigured:
cluster.granular-entry-heal: on
storage.fips-mode-rchecksum: on
transport.address-family: inet
nfs.disable: on
performance.client-io-threads: off
[root@master241 ~]#
6.配置存储卷使用过程中允许某个brick不在线不印象客户端挂载使用(可选操作)
[root@master241 ~]# gluster volume get yinzhengjie-volume cluster.server-quorum-type
Option Value
------ -----
cluster.server-quorum-type off # 默认值为off
[root@master241 ~]#
[root@master241 ~]# gluster volume set yinzhengjie-volume cluster.server-quorum-type none
volume set: success
[root@master241 ~]#
[root@master241 ~]# gluster volume get yinzhengjie-volume cluster.server-quorum-type
Option Value
------ -----
cluster.server-quorum-type none # 将值改为none
[root@master241 ~]#
[root@master241 ~]# gluster volume get yinzhengjie-volume cluster.quorum-type
Option Value
------ -----
cluster.quorum-type auto # 默认值为auto
[root@master241 ~]#
[root@master241 ~]# gluster volume set yinzhengjie-volume cluster.quorum-type none
volume set: success
[root@master241 ~]#
[root@master241 ~]# gluster volume get yinzhengjie-volume cluster.quorum-type
Option Value
------ -----
cluster.quorum-type none # 将值改为none
[root@master241 ~]#
7.配置存储卷支持容量配额功能(可选操作)
[root@master241 ~]# gluster volume quota yinzhengjie-volume enable # 启用容量配额功能
volume quota : success
[root@master241 ~]#
[root@master241 ~]# gluster volume quota yinzhengjie-volume limit-usage / 20GB # 客户端只能使用20GB
volume quota : success
[root@master241 ~]#
8.再次查看卷信息
[root@master241 ~]# gluster volume info yinzhengjie-volume
Volume Name: yinzhengjie-volume
Type: Replicate
Volume ID: f962bd43-e4d4-424b-b693-52714afb6aed
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x 3 = 3
Transport-type: tcp
Bricks:
Brick1: 10.0.0.241:/data/glusterfs/r1
Brick2: 10.0.0.242:/data/glusterfs/r2
Brick3: 10.0.0.243:/data/glusterfs/r3
Options Reconfigured:
features.quota-deem-statfs: on
features.inode-quota: on
features.quota: on # 开启了资源限额功能
cluster.quorum-type: none # 参数生效了
cluster.server-quorum-type: none # 参数生效了
cluster.granular-entry-heal: on
storage.fips-mode-rchecksum: on
transport.address-family: inet
nfs.disable: on
performance.client-io-threads: off
[root@master241 ~]#
参考链接:
https://docs.gluster.org/en/latest/Administrator-Guide/Setting-Up-Volumes/
https://docs.gluster.org/en/latest/Install-Guide/Configure/#set-up-a-gluster-volume
3.GlusterFS客户端挂载
1.加载fuse内核模块
modinfo fuse
modprobe fuse
dmesg | grep -i fuse
2.客户端创建挂载点
[root@master241 ~]# mkdir /k8s-glusterfs
[root@worker242 ~]# mkdir /k8s-glusterfs
[root@worker243 ~]# mkdir /k8s-glusterfs
3.客户端挂着并写入测试数据
[root@master241 ~]# mount -t glusterfs 10.0.0.243:/yinzhengjie-volume /k8s-glusterfs/
[root@master241 ~]#
[root@master241 ~]# df -h | grep k8s-glusterfs
10.0.0.243:/yinzhengjie-volume 20G 0 20G 0% /k8s-glusterfs # 挂载的大小是20GB
[root@master241 ~]#
[root@master241 ~]# cp /etc/os-release /etc/fstab /k8s-glusterfs
[root@master241 ~]#
[root@master241 ~]# ll /k8s-glusterfs
total 6
drwxr-xr-x 4 root root 126 May 27 09:29 ./
drwxr-xr-x 23 root root 4096 May 27 09:25 ../
-rw-r--r-- 1 root root 700 May 27 09:29 fstab
-rw-r--r-- 1 root root 427 May 27 09:29 os-release
[root@master241 ~]#
4.另一个客户端继续挂载测试是否共享数据
[root@worker242 ~]# mount -t glusterfs 10.0.0.241:/yinzhengjie-volume /k8s-glusterfs/
[root@worker242 ~]#
[root@worker242 ~]# df -h | grep k8s-glusterfs
10.0.0.241:/yinzhengjie-volume 20G 0 20G 0% /k8s-glusterfs # 和master241节点挂载的不同的Glusterfs节点。
[root@worker242 ~]#
[root@worker242 ~]# ll /k8s-glusterfs/
total 6
drwxr-xr-x 4 root root 126 May 27 09:29 ./
drwxr-xr-x 23 root root 4096 May 27 09:25 ../
-rw-r--r-- 1 root root 700 May 27 09:29 fstab
-rw-r--r-- 1 root root 427 May 27 09:29 os-release
[root@worker242 ~]#
[root@worker242 ~]# cp /etc/hosts /k8s-glusterfs/
[root@worker242 ~]#
[root@worker242 ~]# ll /k8s-glusterfs/
total 7
drwxr-xr-x 4 root root 139 May 27 09:30 ./
drwxr-xr-x 23 root root 4096 May 27 09:25 ../
-rw-r--r-- 1 root root 700 May 27 09:29 fstab
-rw-r--r-- 1 root root 226 May 27 09:30 hosts
-rw-r--r-- 1 root root 427 May 27 09:29 os-release
[root@worker242 ~]#
5.验证数据是否共享
[root@master241 ~]# ll /k8s-glusterfs
total 7
drwxr-xr-x 4 root root 139 May 27 09:30 ./
drwxr-xr-x 23 root root 4096 May 27 09:25 ../
-rw-r--r-- 1 root root 700 May 27 09:29 fstab
-rw-r--r-- 1 root root 226 May 27 09:30 hosts
-rw-r--r-- 1 root root 427 May 27 09:29 os-release
[root@master241 ~]#
6.卸载Glusterfs
[root@master241 ~]# df -h | grep k8s-glusterfs
10.0.0.243:/yinzhengjie-volume 20G 0 20G 0% /k8s-glusterfs
[root@master241 ~]#
[root@master241 ~]# umount /k8s-glusterfs
[root@master241 ~]#
[root@master241 ~]# df -h | grep k8s-glusterfs
[root@master241 ~]#
三.k8s 1.25+弃用Glusterfs存储卷
1.编写资源清单
[root@master241 volumes]# cat 01-glusterfs-volume.yaml
apiVersion: v1
kind: Endpoints
metadata:
name: glusterfs-cluster
subsets:
- addresses:
- ip: 10.0.0.241
ports:
- port: 1
- addresses:
- ip: 10.0.0.242
ports:
- port: 1
- addresses:
- ip: 10.0.0.243
ports:
- port: 1
---
apiVersion: v1
kind: Service
metadata:
name: glusterfs-cluster
spec:
ports:
- port: 1
---
apiVersion: v1
kind: Pod
metadata:
name: glusterfs
spec:
volumes:
- name: data
glusterfs:
endpoints: glusterfs-cluster
path: yinzhengjie-volume
readOnly: false
containers:
- name: c1
image: registry.cn-hangzhou.aliyuncs.com/yinzhengjie-k8s/apps:v1
volumeMounts:
- mountPath: "/yinzhengjie-glusterfs"
name: data
[root@master241 volumes]#
[root@master241 volumes]#
2.K8S 1.26+移除GlusterFS的支持
温馨提示:
不拿发现，如果K8S版本过高，则K8S 1.26+不在支持GlusterFS。
1.创建存储卷
[root@master241 volumes]# kubectl apply -f 01-glusterfs-volume.yaml
endpoints/glusterfs-cluster created
service/glusterfs-cluster created
Warning: spec.volumes[0].glusterfs: deprecated in v1.25, non-functional in v1.26+
pod/glusterfs created
[root@master241 volumes]#
[root@master241 volumes]# kubectl get nodes -o wide
NAME STATUS ROLES AGE VERSION INTERNAL-IP EXTERNAL-IP OS-IMAGE KERNEL-VERSION CONTAINER-RUNTIME
master241 Ready control-plane 10d v1.31.9 10.0.0.241 <none> Ubuntu 22.04.4 LTS 5.15.0-140-generic containerd://1.6.36
worker242 Ready <none> 10d v1.31.9 10.0.0.242 <none> Ubuntu 22.04.4 LTS 5.15.0-140-generic containerd://1.6.36
worker243 Ready <none> 10d v1.31.9 10.0.0.243 <none> Ubuntu 22.04.4 LTS 5.15.0-140-generic containerd://1.6.36
[root@master241 volumes]#
2.默认不支持该存储卷
[root@master241 volumes]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
glusterfs 0/1 ContainerCreating 0 4m10s <none> worker243 <none> <none>
[root@master241 volumes]#
[root@master241 volumes]#
[root@master241 volumes]# kubectl describe pod glusterfs
Name: glusterfs
Namespace: default
...
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 4m14s default-scheduler Successfully assigned default/glusterfs to worker243
Warning FailedMount 8s (x20 over 4m14s) kubelet Unable to attach or mount volumes: unmounted volumes=[data], unattached volumes=[], failed to process volumes=[data]: failed to get Plugin from volumeSpec for volume "data" err=no volume plugin matched
[root@master241 volumes]#
3.Heketi是GlusterFS和K8S的桥梁？
Heketi提供了一个RESTful管理接口，可用于管理GlusterFS卷的生命周期。借助Heketi，OpenStack Manila、Kubernetes和OpenShift等云服务可以动态地为GlusterFS卷提供任何支持的持久性类型。
Heketi还支持任意数量的GlusterFS集群，允许云服务提供网络文件存储，而不限于单个GlusterFS群集。
Heketi是为GlusterFS提供RESETFUL的API，相当于给glusterfs和K8S之间驾通了桥梁，k8s集群可以通过heketi提供的RESTFUL API完成对GlusterFS的POV申请和管理。
github地址:
https://github.com/heketi/heketi
不过可惜的是，此存储库由所有者于2023年7月6日存档。它现在是只读的。
本文来自博客园，作者：尹正杰，转载请注明原文链接：https://www.cnblogs.com/yinzhengjie/p/18897690，个人微信: "JasonYin2020"（添加时请备注来源及意图备注，有偿付费）
当你的才华还撑不起你的野心的时候，你就应该静下心来学习。当你的能力还驾驭不了你的目标的时候，你就应该沉下心来历练。问问自己，想要怎样的人生。
浙公网安备 33010602011771号