Ceph Reef(18.2.X)的CephFS高可用集群实战案例
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。
一.cephFS基础知识
1. CephFS概述
RBD提供了远程磁盘挂载的问题，但无法做到多个追加共享一个磁盘，如果有一份数据很多客户端都要读写该怎么办呢？这时CephFS作为文件系统解决方案就派上用场了。
CephFS是POSIX兼容的文件系统，它直接使用Ceph存储集群来存储数据。Ceph文件系统于Ceph块设备，同时提供S3和Swift API的Ceph对象存储或者原生库(librados)的实现机制稍显不同。
如上图所示，CephFS支持内核模块或者fuse方式访问，如果宿主机没有安装ceph模块，则可以考虑使用fuse方式访问。可以通过"modinfo ceph"来检查当前宿主机是否有ceph相关内核模块。
[root@ceph141 ~]# modinfo ceph
filename: /lib/modules/5.15.0-119-generic/kernel/fs/ceph/ceph.ko
license: GPL
description: Ceph filesystem for Linux
author: Patience Warnick <patience@newdream.net>
author: Yehuda Sadeh <yehuda@hq.newdream.net>
author: Sage Weil <sage@newdream.net>
alias: fs-ceph
srcversion: 8ABA5D4086D53861D2D72D1
depends: libceph,fscache,netfs
retpoline: Y
intree: Y
name: ceph
vermagic: 5.15.0-119-generic SMP mod_unload modversions
sig_id: PKCS#7
signer: Build time autogenerated kernel key
sig_key: 5F:22:FC:96:FA:F2:9E:F0:76:8D:52:81:4A:3C:6F:F8:61:08:2C:6D
sig_hashalgo: sha512
signature: 76:9E:6E:70:EF:E0:C3:A7:5F:24:C7:C1:2D:EC:24:B9:EE:84:17:64:
F3:18:EC:8F:59:C3:88:64:96:56:1C:FE:97:CE:91:36:C9:BF:98:4B:
A7:F1:63:A5:C8:31:23:5F:12:43:4F:D7:C0:D9:E7:97:72:33:4D:DA:
04:43:60:F9:91:E7:CD:BC:45:30:E4:5A:E0:1B:20:B3:8B:4A:FD:6D:
9F:C6:B8:75:BF:F4:AD:1B:E4:86:FE:85:34:CA:97:DE:17:45:D4:00:
D2:AA:71:3D:18:84:36:D5:33:31:7F:02:8B:1A:26:49:A2:B3:C0:02:
C8:9A:F6:02:7F:B3:82:9C:83:29:3B:F6:ED:E4:BA:BB:0D:F9:E8:27:
04:05:39:29:C8:33:BD:35:8B:77:29:48:89:8F:E9:DD:55:89:3D:CF:
2D:B8:E7:17:8E:20:F6:EE:7D:5B:6D:9A:92:A4:ED:6A:9D:9B:83:1E:
6B:A1:CA:70:28:43:BB:49:88:12:39:65:01:49:2F:0A:4B:D6:9F:26:
EB:D6:E5:FD:35:85:85:E6:CB:E2:87:D1:D3:9D:C8:8F:23:10:DC:D1:
C2:F0:D8:A7:AF:F6:1A:FD:0A:7C:40:1F:40:7B:08:4F:1A:8F:A9:45:
CC:99:57:DB:5E:35:F1:16:49:3A:1B:EF:B4:E6:4B:95:B5:56:A4:D1:
7E:5E:48:10:EE:BC:5C:1A:A7:B0:F1:2E:20:33:5B:55:24:CA:05:A6:
6B:16:20:AE:24:6A:4E:97:1B:81:E0:A1:03:E0:A9:C2:BD:05:36:E6:
09:33:F5:E5:D6:C2:17:CE:28:20:27:59:61:C3:18:06:DB:CC:0F:90:
81:AC:99:26:E1:93:CE:CF:A5:94:56:41:D9:32:C7:6F:FE:F1:8B:9D:
51:4B:82:FC:8B:90:21:B7:96:94:38:4C:DF:EC:0F:85:E5:90:3B:40:
C7:E3:04:7F:D9:F2:81:43:8C:E9:57:18:88:06:B5:3A:F9:E5:A3:01:
52:BD:E7:36:AF:D5:BD:64:16:C2:2A:3D:61:7B:10:09:35:91:C4:5D:
B6:0A:D0:22:2F:8E:A7:E7:48:EE:27:BA:86:E9:BA:96:27:55:E0:F5:
51:36:B7:B0:11:D6:50:6E:85:92:03:DA:95:0E:92:26:A2:7B:5F:AD:
19:C2:A2:CD:71:F3:D3:72:91:8C:79:3D:85:3B:75:C0:AA:81:0F:93:
49:6E:21:F5:38:F5:03:69:17:75:36:04:A8:81:6A:03:2E:7D:0D:10:
81:24:81:4E:78:93:08:8C:1E:CC:BA:5B:47:3A:08:40:A5:CE:4F:3A:
33:DF:A5:1D:4F:8F:0A:62:2A:A3:62:4B
parm: disable_send_metrics:Enable sending perf metrics to ceph cluster (default: on)
[root@ceph141 ~]#
推荐阅读:
https://docs.ceph.com/en/reef/architecture/#ceph-file-system
https://docs.ceph.com/en/reef/cephfs/#ceph-file-system
2.CephFS架构原理
CephFS需要至少运行一个元数据服务器（MDS）守护进程(ceph-mds)，此进程管理与CephFS上存储文件相关的元数据信息。
MDS虽然称为元数据服务，但是它却不存储任何元数据信息，它存在的目的仅仅是让我们rados集群提供存储接口。
客户端在访问文件接口时，首先链接到MDS上，在MDS到内存里面维持元数据的索引信息，从而间接找到去哪个数据节点读取数据。这一点和HDFS文件系统类似。
3.CephFS和NFS对比
相较于NFS来说，它主要有以下特点优势:
- 1.底层数据冗余的功能，底层的roados提供了基本数据冗余功能，因此不存在NFS的单点故障因素;
- 2.底层roados系统有N个存储节点组成，所以数据的存储可以分散I/O，吞吐量较高;
- 3.底层roados系统有N个存储节点组成，所以ceph提供的扩展性要相当的高;
二.cephFS的一主一从架构
推荐阅读:
https://docs.ceph.com/en/reef/cephfs/createfs/
1.创建两个存储池分别用于存储mds的元数据和数据
[root@ceph141 ~]# ceph osd pool create cephfs_data
pool 'cephfs_data' created
[root@ceph141 ~]#
[root@ceph141 ~]# ceph osd pool create cephfs_metadata
pool 'cephfs_metadata' created
[root@ceph141 ~]#
2.创建一个文件系统，名称为"yinzhengjie-cephfs"
[root@ceph141 ~]# ceph fs new yinzhengjie-cephfs cephfs_metadata cephfs_data
Pool 'cephfs_data' (id '6') has pg autoscale mode 'on' but is not marked as bulk.
Consider setting the flag by running
# ceph osd pool set cephfs_data bulk true
new fs with metadata pool 7 and data pool 6
[root@ceph141 ~]#
3.查看创建的存储池
[root@ceph141 ~]# ceph fs ls
name: yinzhengjie-cephfs, metadata pool: cephfs_metadata, data pools: [cephfs_data ]
[root@ceph141 ~]#
[root@ceph141 ~]# ceph mds stat
yinzhengjie-cephfs:0
[root@ceph141 ~]#
[root@ceph141 ~]# ceph -s
cluster:
id: 3cb12fba-5f6e-11ef-b412-9d303a22b70f
health: HEALTH_ERR
1 filesystem is offline
1 filesystem is online with fewer MDS than max_mds
2 pool(s) do not have an application enabled
services:
mon: 3 daemons, quorum ceph141,ceph142,ceph143 (age 23h)
mgr: ceph141.cwgrgj(active, since 27h), standbys: ceph142.ymuzfe
mds: 0/0 daemons up
osd: 7 osds: 7 up (since 23h), 7 in (since 22h)
data:
volumes: 1/1 healthy
pools: 6 pools, 129 pgs
objects: 50 objects, 22 MiB
usage: 553 MiB used, 3.3 TiB / 3.3 TiB avail
pgs: 129 active+clean
[root@ceph141 ~]#
温馨提示:
- 1.此步骤不难发现，存储池的状态无法正常使用，而且集群是有错误的（HEALTH_ERR），因此我们需要先解决这个问题。
4.应用mds的文件系统
[root@ceph141 ~]# ceph orch apply mds yinzhengjie-cephfs
Scheduled mds.yinzhengjie-cephfs update...
[root@ceph141 ~]#
5.添加多个mds服务器
[root@ceph141 ~]# ceph orch daemon add mds yinzhengjie-cephfs ceph143 # 添加第一个mds
Deployed mds.yinzhengjie-cephfs.ceph143.ykwami on host 'ceph143'
[root@ceph141 ~]#
[root@ceph141 ~]# ceph mds stat
yinzhengjie-cephfs:1 {0=yinzhengjie-cephfs.ceph142.xhqpxn=up:active} 1 up:standby
[root@ceph141 ~]#
[root@ceph141 ~]#
[root@ceph141 ~]# ceph orch daemon add mds yinzhengjie-cephfs ceph141 # 添加第二个mds
Deployed mds.yinzhengjie-cephfs.ceph141.fvbkym on host 'ceph141'
[root@ceph141 ~]#
[root@ceph141 ~]# ceph mds stat
yinzhengjie-cephfs:1 {0=yinzhengjie-cephfs.ceph142.xhqpxn=up:active} 2 up:standby
[root@ceph141 ~]#
6.查看cephFS集群的详细信息
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs # 不难发现目前活跃提供服务是ceph141，备用的是ceph142
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 12 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 96.0k 1076G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph142.vpzcat
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
7.再次查看集群状态
[root@ceph141 ~]# ceph -s
cluster:
id: 3cb12fba-5f6e-11ef-b412-9d303a22b70f
health: HEALTH_WARN
2 pool(s) do not have an application enabled
services:
mon: 3 daemons, quorum ceph141,ceph142,ceph143 (age 24h)
mgr: ceph141.cwgrgj(active, since 27h), standbys: ceph142.ymuzfe
mds: 1/1 daemons up, 1 standby # 注意，一个启用，一个备用。
osd: 7 osds: 7 up (since 23h), 7 in (since 22h)
data:
volumes: 1/1 healthy
pools: 6 pools, 129 pgs
objects: 72 objects, 22 MiB
usage: 553 MiB used, 3.3 TiB / 3.3 TiB avail
pgs: 129 active+clean
[root@ceph141 ~]#
8.cephFS的一主一从架构高可用验证
8.1 查看cephFS集群的详细信息
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs # 不难发现目前活跃提供服务是ceph141，备用的是ceph142
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 12 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 96.0k 1076G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph142.vpzcat
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
8.2 ceph141直接关机
[root@ceph141 ~]# init 0
8.3 查看cephFS集群状态
[root@ceph143 ~]# ceph fs status yinzhengjie-cephfs # 很明显，ceph142接管了。【但是在接管之前，可能会卡一阵【30s】，需要迁移数据！】
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 10 13 12 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 64.0k 1614G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph143 ~]#
三.cephFS配置多主一从架构
1 修改之前的状态
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 10 13 12 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 105k 1076G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph141.fvbkym
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
2 修改mds的数量
[root@ceph141 ~]# ceph fs get yinzhengjie-cephfs | grep max_mds # 查看默认的数量为1，说白了，同时仅有一个对外提供服务
max_mds 1
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs set yinzhengjie-cephfs max_mds 2 # 我们可以让2个mds同时对外提供服务
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs get yinzhengjie-cephfs | grep max_mds
max_mds 2
[root@ceph141 ~]#
3 再次查看cephFS
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 10 13 12 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 177k 1076G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
温馨提示:
此时ceph142和ceph141同时对外提供服务，任何一个mds怪掉，都导致集群不可用！
4 添加一个备用的mds
[root@ceph141 ~]# ceph orch daemon add mds yinzhengjie-cephfs ceph143
Deployed mds.yinzhengjie-cephfs.ceph143.rsarvw on host 'ceph143'
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 10 13 12 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 177k 1076G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph143.rsarvw
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
温馨提示:
此事，ceph142和ceph143任意一个mds挂掉，ceph143会立刻接管保证集群的可用性。
5 验证是否生效
5.1 关掉人一个active的mds节点
[root@ceph142 ~]# init 0
5.2 查看cephFS是否生效
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 1 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active(laggy) yinzhengjie-cephfs.ceph142.vpzcat 0 0 0 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 1
POOL TYPE USED AVAIL
cephfs_metadata metadata 121k 1614G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
[root@ceph141 ~]# ceph -s
cluster:
id: 3cb12fba-5f6e-11ef-b412-9d303a22b70f
health: HEALTH_WARN
insufficient standby MDS daemons available
1/3 mons down, quorum ceph141,ceph143
3 osds down
1 host (3 osds) down
Degraded data redundancy: 91/273 objects degraded (33.333%), 37 pgs degraded, 129 pgs undersized
2 pool(s) do not have an application enabled
services:
mon: 3 daemons, quorum ceph141,ceph143 (age 2m), out of quorum: ceph142
mgr: ceph141.cwgrgj(active, since 107s)
mds: 2/2 daemons up
osd: 7 osds: 4 up (since 2m), 7 in (since 22h)
data:
volumes: 1/1 healthy
pools: 6 pools, 129 pgs
objects: 91 objects, 22 MiB
usage: 258 MiB used, 1.6 TiB / 1.6 TiB avail
pgs: 91/273 objects degraded (33.333%)
92 active+undersized
37 active+undersized+degraded
[root@ceph141 ~]#
温馨提示:
由于我的备用节点ceph143莫名了消失了，此时没有人接管142，因此会被标记为"laggy"延迟。
5.3 为了快速修复问题，只能再将143再重启启动
[root@ceph141 ~]# ceph orch daemon add mds yinzhengjie-cephfs ceph143
Deployed mds.yinzhengjie-cephfs.ceph143.vpvgvx on host 'ceph143'
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 resolve yinzhengjie-cephfs.ceph143.vpvgvx 0 0 0 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 121k 1614G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 1 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph143.vpvgvx Reqs: 0 /s 10 13 12 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 1
POOL TYPE USED AVAIL
cephfs_metadata metadata 121k 1614G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
[root@ceph141 ~]#
温馨提示:
启动后，不难发现ceph143由"resolve"（解决）变为"active"(活跃)，说白了，就是备用节点上位了。
5.4 将ceph142节点拉起，观察是否会抢占
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph143.vpvgvx Reqs: 0 /s 10 13 12 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 192k 1614G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph142.vpzcat
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
温馨提示:
不难发现，ceph142处于备用mds，不会抢占之前的角色。
5.5 但是，过了一段时间，我的ceph143节点默认又消失了，此时ceph142正常上位！
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 0 3 2 0
1 active yinzhengjie-cephfs.ceph141.fvbkym Reqs: 0 /s 10 13 11 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 192k 1076G
cephfs_data data 0 1076G
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
[root@ceph141 ~]# ceph -s
cluster:
id: 3cb12fba-5f6e-11ef-b412-9d303a22b70f
health: HEALTH_WARN
insufficient standby MDS daemons available # 由于我的集群ceph143莫名其妙的挂了，导致没有备用的mds，因此此处会告警！
clock skew detected on mon.ceph142
2 pool(s) do not have an application enabled
services:
mon: 3 daemons, quorum ceph141,ceph142,ceph143 (age 2m)
mgr: ceph141.cwgrgj(active, since 8m), standbys: ceph142.ymuzfe
mds: 2/2 daemons up
osd: 7 osds: 7 up (since 2m), 7 in (since 22h)
data:
volumes: 1/1 healthy
pools: 6 pools, 129 pgs
objects: 91 objects, 22 MiB
usage: 394 MiB used, 3.3 TiB / 3.3 TiB avail
pgs: 129 active+clean
[root@ceph141 ~]#
四.cephfs的客户端之借助内核模块挂载
1.查看集群是否正常工作
[root@ceph141 ~]# ceph fs ls
name: yinzhengjie-cephfs, metadata pool: cephfs_metadata, data pools: [cephfs_data ]
[root@ceph141 ~]#
[root@ceph141 ~]#
[root@ceph141 ~]# ceph fs status yinzhengjie-cephfs
yinzhengjie-cephfs - 0 clients
================
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 active yinzhengjie-cephfs.ceph142.vpzcat Reqs: 0 /s 10 13 12 0
POOL TYPE USED AVAIL
cephfs_metadata metadata 192k 1076G
cephfs_data data 0 1076G
STANDBY MDS
yinzhengjie-cephfs.ceph141.fvbkym
MDS version: ceph version 18.2.4 (e7ad5345525c7aa95470c26863873b581076945d) reef (stable)
[root@ceph141 ~]#
[root@ceph141 ~]# ceph -s
cluster:
id: 3cb12fba-5f6e-11ef-b412-9d303a22b70f
health: HEALTH_OK
services:
mon: 3 daemons, quorum ceph141,ceph142,ceph143 (age 61m)
mgr: ceph141.cwgrgj(active, since 68m), standbys: ceph142.ymuzfe
mds: 1/1 daemons up, 1 standby
osd: 7 osds: 7 up (since 61m), 7 in (since 23h)
data:
volumes: 1/1 healthy
pools: 3 pools, 65 pgs
objects: 43 objects, 455 KiB
usage: 374 MiB used, 3.3 TiB / 3.3 TiB avail
pgs: 65 active+clean
[root@ceph141 ~]#
2.管理节点创建用户并导出钥匙环和key文件
2.1 创建用户并授权
[root@ceph141 ~]# ceph auth add client.yinzhengjiefs mon 'allow r' mds 'allow rw' osd 'allow rwx'
added key for client.yinzhengjiefs
[root@ceph141 ~]#
[root@ceph141 ~]# ceph auth get client.yinzhengjiefs
[client.yinzhengjiefs]
key = AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
caps mds = "allow rw"
caps mon = "allow r"
caps osd = "allow rwx"
[root@ceph141 ~]#
2.2 导出认证信息
[root@ceph141 ~]# ceph auth get client.yinzhengjiefs > ceph.client.yinzhengjiefs.keyring
[root@ceph141 ~]#
[root@ceph141 ~]# cat ceph.client.yinzhengjiefs.keyring
[client.yinzhengjiefs]
key = AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
caps mds = "allow rw"
caps mon = "allow r"
caps osd = "allow rwx"
[root@ceph141 ~]#
[root@ceph141 ~]# ceph auth print-key client.yinzhengjiefs > yinzhengjiefs.key
[root@ceph141 ~]#
[root@ceph141 ~]# more yinzhengjiefs.key
AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
[root@ceph141 ~]#
3.将钥匙环和秘钥key拷贝到客户端指定目录
[root@ceph141 ~]# scp ceph.client.yinzhengjiefs.keyring yinzhengjiefs.key ceph143:/etc/ceph/
[root@ceph141 ~]#
[root@ceph141 ~]# scp ceph.client.yinzhengjiefs.keyring ceph142:/etc/ceph/
4.基于secretfile进行挂载
4.1 查看本地文件
[root@ceph143 ~]# ll /etc/ceph/ | grep yinzhengjie
-rw-r--r-- 1 root root 140 Aug 22 16:49 ceph.client.yinzhengjiefs.keyring
-rw-r--r-- 1 root root 40 Aug 22 16:49 yinzhengjiefs.key
[root@ceph143 ~]#
4.2 查看本地解析记录
[root@ceph143 ~]# grep ceph /etc/hosts
10.0.0.141 ceph141
10.0.0.142 ceph142
10.0.0.143 ceph143
[root@ceph143 ~]#
4.3 基于key文件进行挂载并尝试写入数据
[root@ceph143 ~]# df -h | grep mnt
[root@ceph143 ~]#
[root@ceph143 ~]# mount -t ceph ceph141:6789,ceph142:6789,ceph143:6789:/ /mnt -o name=yinzhengjiefs,secretfile=/etc/ceph/yinzhengjiefs.key
[root@ceph143 ~]#
[root@ceph143 ~]# df -h | grep mnt
10.0.0.141:6789,10.0.0.142:6789,10.0.0.143:6789:/ 1.1T 0 1.1T 0% /mnt
[root@ceph143 ~]#
[root@ceph143 ~]# ll /mnt/
total 4
drwxr-xr-x 2 root root 0 Aug 22 15:12 ./
drwxr-xr-x 21 root root 4096 Aug 21 17:52 ../
[root@ceph143 ~]#
[root@ceph143 ~]# cp /etc/os-release /etc/fstab /etc/hosts /mnt/
[root@ceph143 ~]#
[root@ceph143 ~]# ll /mnt/
total 6
drwxr-xr-x 2 root root 3 Aug 22 16:52 ./
drwxr-xr-x 21 root root 4096 Aug 21 17:52 ../
-rw-r--r-- 1 root root 657 Aug 22 16:52 fstab
-rw-r--r-- 1 root root 283 Aug 22 16:52 hosts
-rw-r--r-- 1 root root 386 Aug 22 16:52 os-release
[root@ceph143 ~]#
[root@ceph143 ~]# umount /mnt
[root@ceph143 ~]#
[root@ceph143 ~]# ll /mnt/
total 8
drwxr-xr-x 2 root root 4096 Aug 10 2023 ./
drwxr-xr-x 21 root root 4096 Aug 21 17:52 ../
[root@ceph143 ~]#
5.基于KEY进行挂载，无需拷贝秘钥文件！
[root@ceph143 ~]# more /etc/ceph/yinzhengjiefs.key
AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
[root@ceph143 ~]#
[root@ceph143 ~]# df -h | grep mnt
[root@ceph143 ~]#
[root@ceph143 ~]# mount -t ceph ceph141:6789,ceph142:6789,ceph143:6789:/ /mnt -o name=yinzhengjiefs,secret=AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
[root@ceph143 ~]#
[root@ceph143 ~]# df -h | grep mnt
10.0.0.141:6789,10.0.0.142:6789,10.0.0.143:6789:/ 1.1T 0 1.1T 0% /mnt
[root@ceph143 ~]#
[root@ceph143 ~]#
[root@ceph143 ~]# cp /etc/netplan/00-installer-config.yaml /mnt/
[root@ceph143 ~]#
[root@ceph143 ~]# ll /mnt/
total 7
drwxr-xr-x 2 root root 4 Aug 22 16:56 ./
drwxr-xr-x 21 root root 4096 Aug 21 17:52 ../
-rw-r--r-- 1 root root 237 Aug 22 16:56 00-installer-config.yaml
-rw-r--r-- 1 root root 657 Aug 22 16:52 fstab
-rw-r--r-- 1 root root 283 Aug 22 16:52 hosts
-rw-r--r-- 1 root root 386 Aug 22 16:52 os-release
[root@ceph143 ~]#
6.在启动一个客户端测试
[root@ceph142 ~]# df -h | grep mnt
[root@ceph142 ~]#
[root@ceph142 ~]# mount -t ceph ceph141:6789,ceph142:6789,ceph143:6789:/ /mnt -o name=yinzhengjiefs,secret=AQAF+8ZmIc85CRAABc/93tPRlyJFamvBR6oCOw==
[root@ceph142 ~]#
[root@ceph142 ~]# df -h | grep mnt
10.0.0.141:6789,10.0.0.142:6789,10.0.0.143:6789:/ 1.1T 0 1.1T 0% /mnt
[root@ceph142 ~]#
[root@ceph142 ~]# cp /etc/hostname /mnt/
[root@ceph142 ~]#
[root@ceph142 ~]# ll /mnt/
total 7
drwxr-xr-x 2 root root 5 Aug 22 17:00 ./
drwxr-xr-x 21 root root 4096 Aug 21 17:42 ../
-rw-r--r-- 1 root root 237 Aug 22 16:56 00-installer-config.yaml
-rw-r--r-- 1 root root 657 Aug 22 16:52 fstab
-rw-r--r-- 1 root root 8 Aug 22 17:00 hostname
-rw-r--r-- 1 root root 283 Aug 22 16:52 hosts
-rw-r--r-- 1 root root 386 Aug 22 16:52 os-release
[root@ceph142 ~]#
7.cephfs配置开机自动挂载
推挤阅读:
https://www.cnblogs.com/yinzhengjie/p/14305987.html#二cephfs开机自动挂载的三种方式
五.cephfs的客户端之基于用户空间fuse方式访问
1.FUSE概述
对于某些操作系统来说，它没有提供对应的ceph内核模块，我们还需要使用CephFS的话，可以通过FUSE方式来实现。
FUSE英文全称为:"Filesystem in Userspace"，用于非特权用户能够无需操作内核而创建文件系统，但需要单独安装"ceph-fuse"程序包。
2.FUSE实战案例
2.1 安装ceph-fuse程序包
[root@ceph143 ~]# apt -y install ceph-fuse ceph-common
2.2 使用ceph-fuse工具挂载cephFS
1.创建挂载点
[root@ceph143 ~]# mkdir /yinzhengjie/cephfs
2.开始挂载ceph
[root@ceph143 ~]# ll /etc/ceph/
total 32
drwxr-xr-x 2 root root 4096 Aug 28 21:35 ./
drwxr-xr-x 132 root root 12288 Aug 28 21:32 ../
-rw------- 1 root root 151 Aug 28 21:35 ceph.client.admin.keyring
-rw-r--r-- 1 root root 259 Aug 28 21:35 ceph.conf
-rw-r--r-- 1 root root 595 Aug 28 21:35 ceph.pub
-rw-r--r-- 1 root root 92 Feb 8 2024 rbdmap
[root@ceph143 ~]#
[root@ceph143 ~]# ceph-fuse -n client.admin -m ceph141:6789,ceph142:6789,ceph143:6789 /yinzhengjie/cephfs/
3.验证是否挂载成功
[root@ceph143 ~]# mount | grep cephfs
ceph-fuse on /yinzhengjie/cephfs type fuse.ceph-fuse (rw,nosuid,nodev,relatime,user_id=0,group_id=0,allow_other)
[root@ceph143 ~]#
2.3 查看挂载效果
[root@ceph143 ~]# stat -f /yinzhengjie/cephfs/
File: "/yinzhengjie/cephfs/"
ID: 0 Namelen: 255 Type: fuseblk
Block size: 4194304 Fundamental block size: 4194304
Blocks: Total: 121553 Free: 121553 Available: 121553
Inodes: Total: 1 Free: -1
[root@ceph143 ~]#
2.4 验证挂载效果
[root@ceph143 ~]# ll /yinzhengjie/cephfs/
total 5
drwxr-xr-x 2 root root 0 Aug 31 17:01 ./
drwxr-xr-x 4 root root 4096 Aug 31 16:56 ../
[root@ceph143 ~]#
[root@ceph143 ~]# cp /etc/os-release /yinzhengjie/cephfs/
[root@ceph143 ~]#
[root@ceph143 ~]# ll /yinzhengjie/cephfs/
total 5
drwxr-xr-x 2 root root 0 Aug 31 17:02 ./
drwxr-xr-x 4 root root 4096 Aug 31 16:56 ../
-rw-r--r-- 1 root root 386 Aug 31 17:02 os-release
[root@ceph143 ~]#
[root@ceph143 ~]# mount | grep cephfs
ceph-fuse on /yinzhengjie/cephfs type fuse.ceph-fuse (rw,nosuid,nodev,relatime,user_id=0,group_id=0,allow_other)
[root@ceph143 ~]#
2.5 卸载cephfs
[root@ceph143 ~]# mount | grep cephfs
ceph-fuse on /yinzhengjie/cephfs type fuse.ceph-fuse (rw,nosuid,nodev,relatime,user_id=0,group_id=0,allow_other)
[root@ceph143 ~]#
[root@ceph143 ~]# umount /yinzhengjie/cephfs
[root@ceph143 ~]#
[root@ceph143 ~]# mount | grep cephfs
[root@ceph143 ~]#
3.配置基于fuse方式的开机启动挂载
[root@ceph143 ~]# mount | grep cephfs
[root@ceph143 ~]#
[root@ceph143 ~]# echo 'none /yinzhengjie/cephfs fuse.ceph ceph.id=admin,ceph.conf=/etc/ceph/ceph.conf,_netdev,noatime 0 0' >> /etc/fstab
[root@ceph143 ~]#
[root@ceph143 ~]# mount -a
2024-08-31T17:08:09.741+0800 7f64ca2f23c0 -1 init, newargv = 0x557852fc6050 newargc=15
2024-08-31T17:08:09.741+0800 7f64ca2f23c0 -1 init, args.argv = 0x557853108800 args.argc=4
ceph-fuse[86672]: starting ceph client
ceph-fuse[86672]: starting fuse
[root@ceph143 ~]#
[root@ceph143 ~]# mount | grep cephfs
ceph-fuse on /yinzhengjie/cephfs type fuse.ceph-fuse (rw,nosuid,nodev,noatime,user_id=0,group_id=0,allow_other)
[root@ceph143 ~]#
[root@ceph143 ~]# ls -l /yinzhengjie/cephfs/
total 1
-rw-r--r-- 1 root root 386 Aug 31 17:02 os-release
[root@ceph143 ~]#
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。
欢迎交流学习技术交流，个人微信: "JasonYin2020"（添加时请备注来源及意图备注）
作者: 尹正杰, 博客: https://www.cnblogs.com/yinzhengjie/p/18375087
