本文详细记录了Ceph集群的搭建过程，包括对象存储、块设备和文件系统的创建与验证。通过模拟osd节点故障和恢复，测试了不同副本丢失情况下集群的稳定性和数据恢复能力。此外，还涉及到了监控节点损失的情况以及集群在更换IP后的恢复验证。测试结果显示，Ceph在多种故障场景下能有效保护数据并恢复正常服务。
验证剧本
- ceph环境搭建，需要搭建14版本的ceph集群，提供三种数据接口，1.对象存储 2.块设备 3.文件系统
- 数据导入，将原来的ceph数据（对象存储）导入到新的ceph集群中
- osd节点灾备恢复，分别对三种数据接口，进行osd节点摘除，按照三副本进行：
情形一：按照一个文件的分布，按照摘除1个osd
结论：摘除一个节点后，可以继续进行文件的读写；三个副本恢复后，仍然可以进行读写
情形二：摘除2个osd节点
结论：摘除两个节点后，不能继续对文件进行读写，只能读取数据，三个副本恢复后，可以正常进行读写 - 更换ip后，验证ceph集群的恢复
结论：更换ip后，步骤3的三种数据接口，仍然可以正常读取文件，且文件md5值一致 - ceph节点损失恢复，按照5个监控节点，验证损失1个节点和三个节点集群的恢复情况
情形一：监控节点损失1到2个节点
结论：摘除两个节点后，ceph仍然能够正常对外提供三种数据接口
情形二：监控节点损失3个节点
结论：ceph不能正常对外提供数据接口，且恢复一个节点后，ceph能够正常对外提供三种数据接口
集群搭建
参加：ceph集群搭建过程
创建文件接口
创建块设备
参加：ceph集群搭建过程 之 挂载块设备 过程
创建ceph fs文件系统
验证
块设备
[root@ceph101 test_rbd]# rados -p test_rbd ls
rbd_data.2836b2f53ea86.000000000000042d
rbd_data.2836b2f53ea86.0000000000000436
rbd_data.2836b2f53ea86.0000000000000421
rbd_data.2836b2f53ea86.0000000000000800
rbd_object_map.2836b2f53ea86
rbd_data.2836b2f53ea86.0000000000000429
rbd_data.2836b2f53ea86.0000000000000439
rbd_data.2836b2f53ea86.0000000000000060
rbd_data.2836b2f53ea86.00000000000009e1
rbd_data.2836b2f53ea86.000000000000042b
rbd_data.2836b2f53ea86.000000000000043d
rbd_data.2836b2f53ea86.0000000000000427
rbd_data.2836b2f53ea86.000000000000042c
rbd_directory
rbd_header.2836b2f53ea86
rbd_data.2836b2f53ea86.0000000000000422
rbd_data.2836b2f53ea86.0000000000000423
rbd_data.2836b2f53ea86.00000000000000a0
rbd_data.2836b2f53ea86.0000000000000430
rbd_data.2836b2f53ea86.000000000000042f
rbd_info
rbd_data.2836b2f53ea86.0000000000000020
rbd_data.2836b2f53ea86.000000000000043c
rbd_data.2836b2f53ea86.0000000000000420
rbd_data.2836b2f53ea86.0000000000000320
rbd_data.2836b2f53ea86.00000000000009ff
rbd_data.2836b2f53ea86.0000000000000437
rbd_data.2836b2f53ea86.0000000000000360
rbd_data.2836b2f53ea86.000000000000042a
rbd_data.2836b2f53ea86.0000000000000200
rbd_data.2836b2f53ea86.0000000000000433
rbd_data.2836b2f53ea86.0000000000000600
rbd_data.2836b2f53ea86.000000000000043a
rbd_data.2836b2f53ea86.0000000000000620
rbd_data.2836b2f53ea86.000000000000043b
rbd_data.2836b2f53ea86.0000000000000434
rbd_data.2836b2f53ea86.000000000000043e
rbd_data.2836b2f53ea86.00000000000000e0
rbd_data.2836b2f53ea86.0000000000000438
rbd_data.2836b2f53ea86.0000000000000425
rbd_data.2836b2f53ea86.0000000000000120
rbd_data.2836b2f53ea86.00000000000009e0
rbd_data.2836b2f53ea86.0000000000000400
rbd_data.2836b2f53ea86.0000000000000001
rbd_data.2836b2f53ea86.0000000000000428
rbd_data.2836b2f53ea86.00000000000009c1
rbd_data.2836b2f53ea86.0000000000000432
rbd_data.2836b2f53ea86.0000000000000426
rbd_data.2836b2f53ea86.00000000000009c0
rbd_data.2836b2f53ea86.0000000000000431
rbd_data.2836b2f53ea86.0000000000000435
rbd_id.test_rbd_image_1
rbd_data.2836b2f53ea86.000000000000042e
rbd_data.2836b2f53ea86.000000000000043f
rbd_data.2836b2f53ea86.0000000000000009
rbd_data.2836b2f53ea86.0000000000000424
rbd_data.2836b2f53ea86.0000000000000000
[root@ceph101 test_rbd]# ceph osd map test_rbd rbd_id.test_rbd_image_1
osdmap e3035 pool 'test_rbd' (43) object 'rbd_id.test_rbd_image_1' -> pg 43.ce44bb37 (43.17) -> up ([3,1,8], p3) acting ([3,1,8], p3)
[root@ceph101 ~]# ceph osd tree
ID CLASS WEIGHT TYPE NAME STATUS REWEIGHT PRI-AFF
-1 5.85956 root default
-3 0.97659 host ceph101
0 hdd 0.48830 osd.0 up 1.00000 1.00000
1 hdd 0.48830 osd.1 up 1.00000 1.00000
-5 0.97659 host ceph102
2 hdd 0.48830 osd.2 up 1.00000 1.00000
3 hdd 0.48830 osd.3 up 1.00000 1.00000
-7 0.97659 host ceph103
4 hdd 0.48830 osd.4 up 1.00000 1.00000
5 hdd 0.48830 osd.5 up 1.00000 1.00000
-9 0.97659 host ceph104
6 hdd 0.48830 osd.6 up 1.00000 1.00000
7 hdd 0.48830 osd.7 up 1.00000 1.00000
-11 0.97659 host ceph105
8 hdd 0.48830 osd.8 up 1.00000 1.00000
9 hdd 0.48830 osd.9 up 1.00000 1.00000
-13 0.97659 host ceph106
10 hdd 0.48830 osd.10 up 1.00000 1.00000
11 hdd 0.48830 osd.11 up 1.00000 1.00000
#查看pg分布（全局）
[root@ceph101 ~]# ceph pg ls
##或者按照osd查看
[root@ceph101 ~]# ceph pg ls-by-osd osd.0
##更精确捕获
[root@ceph101 ~]#ceph pg ls |awk '{printf"%s\t%s\n",$1,$15}'
PG OBJECTS DEGRADED MISPLACED UNFOUND BYTES OMAP_BYTES* OMAP_KEYS* LOG STATE SINCE VERSION REPORTED UP ACTING SCRUB_STAMP DEEP_SCRUB_STAMP
1.0 0 0 0 0 0 0 0 0 active+clean 4h 0'0 3046:3060 [7,5,2]p7 [7,5,2]p7 2022-05-09T06:30:28.561204+0000 2022-05-05T20:49:31.217124+0000
24.0 0 0 0 0 0 0 0 0 active+clean 4h 0'0 3046:1865 [9,0,6]p9 [9,0,6]p9 2022-05-09T21:57:21.361105+0000 2022-05-04T21:21:44.529734+0000
24.1 2 0 0 0 1134 0 0 24 active+clean 4h 2610'24 3046:1898 [3,9,4]p3 [3,9,4]p3 2022-05-10T05:08:18.923492+0000 2022-05-09T02:37:15.839476+0000
24.2 1 0 0 0 939 0 0 13 active+clean 4h 2610'13
..................
43.17 4 0 0 0 12582929 0 0 4 active+clean 4h 3034'4 3046:77 [3,1,8]p3 [3,1,8]p3 2022-05-10T02:36:03.721277+0000 2022-05-10T02:36:03.721277+0000
.....................
ceph副本验证
块对象
[root@ceph101 ~]# ceph osd pool create test_rbd_1 32
pool 'test_rbd_1' created
[root@ceph101 ~]# ceph osd pool application enable test_rbd_1 rbd
enabled application 'rbd' on pool 'test_rbd_1'
[root@ceph101 ~]# rados -p test_rbd_1 put bcsj.note caoyong/bcsj.txt
[root@ceph101 ~]# rados -p test_rbd_1 ls
bcsj.note
[root@ceph101 ~]# ceph osd map test_rbd bcsj.txt
osdmap e3048 pool 'test_rbd' (43) object 'bcsj.txt' -> pg 43.c7ddb025 (43.5) -> up ([11,9,4], p11) acting ([11,9,4], p11)
#禁用副本复制功能
[root@ceph101 test_rbd]# ceph osd set noout
noout is set
[root@ceph101 test_rbd]# ceph osd set norebalance
norebalance is set
[root@ceph101 test_rbd]# ceph osd set norecover
norecover is set
[root@ceph101 test_rbd]# ceph osd set nobackfill
nobackfill is set
[root@ceph101 test_rbd]# rados -p test_rbd append bcsj.txt append.txt
[root@ceph101 test_rbd]# rados -p test_rbd get bcsj.txt bcsj.txt.download
[root@ceph101 test_rbd]# ls
append.txt bcsj.txt.download
[root@ceph101 test_rbd]# tail -10 bcsj.txt.download
就如同夜幕降临，白日西沉。
更多精彩好书，更多原创手机电子书，请登陆奇书网--Www.Qisuu.Com1
[root@ceph101 test_rbd]# ceph orch daemon rm osd.9 --force
Removed osd.9 from host 'ceph105'
[root@ceph101 test_rbd]# ceph osd rm osd.9
removed osd.9
[root@ceph101 test_rbd]# ceph osd out osd.9
osd.9 does not exist.
#两个副本，可以正常对文件进行读写
[root@ceph101 test_rbd]# ceph osd map test_rbd bcsj.txt
osdmap e3054 pool 'test_rbd' (43) object 'bcsj.txt' -> pg 43.c7ddb025 (43.5) -> up ([11,4], p11) acting ([11,4], p11)
[root@ceph101 test_rbd]# cat append.txt
append 2 copies
---------------2-------copies-------------
[root@ceph101 test_rbd]# rados -p test_rbd append bcsj.txt append.txt
[root@ceph101 test_rbd]# rados -p test_rbd get bcsj.txt bcsj.txt.2.copies
[root@ceph101 test_rbd]# tail -10 bcsj.txt.2.copies
更多精彩好书，更多原创手机电子书，请登陆奇书网--Www.Qisuu.Com1
append 2 copies
---------------2-------copies-------------
[root@ceph101 test_rbd]# ceph orch daemon rm osd.4 --force
Removed osd.4 from host 'ceph103'
#剩下一副本后，仍然可以进行读写
[root@ceph101 test_rbd]# ceph osd map test_rbd bcsj.txt
osdmap e3056 pool 'test_rbd' (43) object 'bcsj.txt' -> pg 43.c7ddb025 (43.5) -> up ([11], p11) acting ([11], p11)
[root@ceph101 test_rbd]# ceph osd out osd.4
marked out osd.4.
[root@ceph101 test_rbd]# ceph osd rm osd.4
removed osd.4
[root@ceph101 test_rbd]# cat append.txt
append 1 copies
---------------1-------copies-------------
[root@ceph101 test_rbd]# rados -p test_rbd append bcsj.txt append.txt
[root@ceph101 test_rbd]# rados -p test_rbd get bcsj.txt bcsj.txt.1.copies
[root@ceph101 test_rbd]# tail -12 bcsj.txt.1.copies
更多精彩好书，更多原创手机电子书，请登陆奇书网--Www.Qisuu.Com1
append 2 copies
---------------2-------copies-------------
append 1 copies
---------------1-------copies-------------
#恢复集群
[root@ceph101 ~]# ceph osd set noout
noout is set
[root@ceph101 ~]# ceph osd set norebalance
norebalance is set
[root@ceph101 ~]# ceph osd set norecover
norecover is set
[root@ceph101 ~]# ceph osd set nobackfill
nobackfill is set
#将其他osd重新加入，可以看到文件重新恢复成三副本
[root@ceph101 test_rbd]# ceph osd map test_rbd bcsj.txt
osdmap e3048 pool 'test_rbd' (43) object 'bcsj.txt' -> pg 43.c7ddb025 (43.5) -> up ([11,9,4], p11) acting ([11,9,4], p11)
对象网关
参照快设备，可以找到对象网关接口下，副本的分布情况
[root@ceph1 ~]# ceph osd lspools
1 device_health_metrics
2 .rgw.root
43 knowdee
54 cp_pool
57 zone_02.rgw.log
58 zone_03.rgw.log
59 zone_02.rgw.meta
60 zone_02.rgw.control
61 zone_02.rgw.buckets.index
64 zone_02.rgw.buckets.data
[root@ceph1 ~]# radosgw-admin bucket stats --bucket caoyong
{
"bucket": "caoyong",
"num_shards": 11,
"tenant": "",
"zonegroup": "9eadce15-982d-49b2-8b5f-b640ab920e82",
"placement_rule": "default-placement",
"explicit_placement": {
"data_pool": "",
"data_extra_pool": "",
"index_pool": ""
},
"id": "b80f8394-1f06-47e9-b2a0-bc6cfd2fc26a.134229.1",
"marker": "b80f8394-1f06-47e9-b2a0-bc6cfd2fc26a.134229.1",
"index_type": "Normal",
"owner": "rgw_admin",
"ver": "0#1,1#1,2#3,3#1,4#1,5#1,6#1,7#1,8#1,9#1,10#1",
"master_ver": "0#0,1#0,2#0,3#0,4#0,5#0,6#0,7#0,8#0,9#0,10#0",
"mtime": "2022-04-27T08:50:27.397030Z",
"creation_time": "2022-04-27T08:50:22.362309Z",
"max_marker": "0#,1#,2#00000000002.14.5,3#,4#,5#,6#,7#,8#,9#,10#",
"usage": {
"rgw.main": {
"size": 19098141,
"size_actual": 19099648,
"size_utilized": 19098141,
"size_kb": 18651,
"size_kb_actual": 18652,
"size_kb_utilized": 18651,
"num_objects": 2
}
},
"bucket_quota": {
"enabled": false,
"check_on_raw": false,
"max_size": -1,
"max_size_kb": 0,
"max_objects": -1
}
}
[root@ceph1 ~]# ceph osd map zone_02.rgw.buckets.index .dir.b80f8394-1f06-47e9-b2a0-bc6cfd2fc26a.134229.1
osdmap e4183 pool 'zone_02.rgw.buckets.index' (61) object '.dir.b80f8394-1f06-47e9-b2a0-bc6cfd2fc26a.134229.1' -> pg 61.aaf566ea (61.2) -> up ([0,1,6], p0) acting ([0,1,6], p0)
cephfs文件系统
参照快设备，可以找到cephfs文件系统接口下，副本的分布情况
[root@ceph1 ~]# ceph osd pool create cephfs_data 32
pool 'cephfs_data' created
[root@ceph1 ~]# ceph osd pool create cephfs_metadata 32
pool 'cephfs_metadata' created
[root@ceph1 ~]# ceph fs new cephfs cephfs_metadata cephfs_data
[root@ceph1 ~]# ceph fs volume create cephfs
new fs with metadata pool 68 and data pool 67
[root@ceph1 ~]# ceph fs status
cephfs - 0 clients
======
RANK STATE MDS ACTIVITY DNS INOS
0 active cephfs.ceph6.cqdgnc Reqs: 0 /s 10 13
POOL TYPE USED AVAIL
cephfs.cephfs.meta metadata 1536k 1895G
cephfs.cephfs.data data 0 1895G
STANDBY MDS
cephfs.ceph2.afksaf
MDS version: ceph version 15.2.16 (d46a73d6d0a67a79558054a3a5a72cb561724974) octopus (stable)
[root@ceph1 ~]# ceph auth get-or-create client.cephfs mon 'allow *' mds 'allow *' osd 'allow * pool=cephfs_data, allow * pool=cephfs_metadata'
[client.cephfs]
key = AQAG9oFimrR/ORAA0mOMbh8uknrRLrA0DD/SCA==
[root@ceph1 ~]# ceph auth get client.cephfs
exported keyring for client.cephfs
[client.cephfs]
key = AQAG9oFimrR/ORAA0mOMbh8uknrRLrA0DD/SCA==
caps mds = "allow rw"
caps mon = "allow r"
caps osd = "allow rw pool=cephfs_data, allow rw pool=cephfs_metadata"
[root@ceph1 ~]# pwd
/root
[root@ceph1 ~]# mkdir cephfs
[root@ceph1 ~]# mount -t ceph 172.70.10.161:6789:/ /root/cephfs -o name=cephfs,secret=AQAG9oFimrR/ORAA0mOMbh8uknrRLrA0DD/SCA==
[root@ceph1 ~]# df -h
文件系统 容量 已用 可用 已用% 挂载点
devtmpfs 7.9G 0 7.9G 0% /dev
tmpfs 7.9G 0 7.9G 0% /dev/shm
tmpfs 7.9G 860M 7.0G 11% /run
tmpfs 7.9G 0 7.9G 0% /sys/fs/cgroup
/dev/mapper/centos-root 50G 17G 34G 34% /
/dev/vda1 1014M 176M 839M 18% /boot
/dev/mapper/centos-home 441G 33M 441G 1% /home
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/29b83745237d0b8c1998f9e8b8a4e237d2851c8fdc3085358a72e3d4279bdda2/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/8164d8132289057cfc142839d15a4ef22e2d6e39e89a93f610786a2469e23cd5/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/73d1c01c8bbf1c88189540fdb06a6c1ede1a52720925e6c0fc5bd4d94c9fa960/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/87c970108e45f919a9f2e6d6634f09747cfaffb68e2ce6fe90fa7997cf8abc33/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/4d5db84cd24c09eaf39d2dde96639a87c4053cae625bc99fe431326052e95d0f/merged
tmpfs 1.6G 0 1.6G 0% /run/user/0
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/875dad643dce157f200a03d9c4d8b0ddcdaef834c7df971d8976312a95b68871/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/3a9b6217e2acc2536547dcbf8501766469188fc33af82dfbe08bae7d06a9cc2b/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/97b51af97503d53caf12b3a0bc5aa96c0b8e664be225c201302f093d62caa429/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/2744f55d60cd71d70df6b99ce870a25e41024465b2deeea38fba93f0f831c9ed/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/9ed5a8d1866530c548c469961413b9d9c4ac86328cff32ed9382847be5916ea1/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/cc51e991ec059837464d744d9d8f36fbca8e13d5a30b36cbdc52e04993d16593/merged
overlay 50G 17G 34G 34% /var/lib/docker/overlay2/308e2d063bab1c5c97bf6796da05f7ed83c4adc4c846a1a6c232f31fccef70c6/merged
172.70.10.161:6789:/ 1.9T 0 1.9T 0% /root/cephfs
[root@ceph1 cephfs]# pwd
/root/cephfs
[root@ceph1 cephfs]# cat caoyong.txt
iii
[root@ceph1 cephfs]# rados ls -p cephfs_data
10000000004.00000000
[root@ceph1 cephfs]# ceph osd map cephfs_data 10000000004.00000000
osdmap e4202 pool 'cephfs_data' (67) object '10000000004.00000000' -> pg 67.17c5e6c5 (67.5) -> up ([0,4,10], p0) acting ([0,4,10], p0)
[root@ceph1 cephfs]# rados -p cephfs_data get 10000000004.00000000 bcsj.txt.2.copies
[root@ceph1 cephfs]# ll
总用量 1
-rw-r--r-- 1 root root 4 5月 16 15:03 bcsj.txt.2.copies
-rw-r--r-- 1 root root 4 5月 16 14:59 caoyong.txt
[root@ceph1 cephfs]# md5sum bcsj.txt.2.copies
3b7c423ba4e937ec11b6f7a7814b9960 bcsj.txt.2.copies
[root@ceph1 cephfs]# md5sum caoyong.txt
3b7c423ba4e937ec11b6f7a7814b9960 caoyong.txt
[root@ceph1 cephfs]# vim bcsj.txt.2.copies
其他的常用命令
config set osd osd_pool_default_min_size 1
ceph osd set noout
ceph osd set norebalance
ceph osd set norecover
ceph osd set nobackfill
ceph osd pool get test_rbd min_size
ceph orch device zap ceph102 /dev/vdb --force
ceph orch daemon add osd ceph102:/dev/vdb
ceph orch daemon rm osd.6 --force
ceph osd out osd.6
ceph osd rm osd.6
rados -p test_rbd_1 append bcsj.txt append.txt
ceph osd pool get test_rbd_1 min_size
ceph osd pool delete test_rbd
[root@ceph101 ~]# ceph pg map 24.1
osdmap e3304 pg 24.1 (24.1) -> up [3,9,4] acting [3,9,4]
#设置min_size = 1
$ bin/ceph osd pool set test_pool min_size 1
set pool 1 min_size to 1
ceph osd pool get test_pool min_size
集群不健康：
现象一
[root@ceph1 ~]# ceph health detail
HEALTH_WARN 2 failed cephadm daemon(s); 3 daemons have recently crashed
[WRN] CEPHADM_FAILED_DAEMON: 2 failed cephadm daemon(s)
daemon rgw.realm_1.zone_02.ceph1.bnqyza on ceph1 is in unknown state
daemon rgw.realm_1.zone_03.ceph1.liwnmb on ceph1 is in unknown state
[WRN] RECENT_CRASH: 3 daemons have recently crashed
client.rgw.realm_1.zone_02.ceph1.bnqyza crashed on host ceph1 at 2022-05-13T12:36:17.438998Z
client.rgw.realm_1.zone_03.ceph1.liwnmb crashed on host ceph1 at 2022-05-13T12:38:12.049014Z
client.rgw.realm_1.zone_03.ceph1.liwnmb crashed on host ceph1 at 2022-05-13T12:38:12.324601Z
解决：
#删除无用的service
[root@ceph1 ~]# ceph orch daemon rm rgw.realm_1.zone_02.ceph1.bnqyza --force
Removed rgw.realm_1.zone_02.ceph1.bnqyza from host 'ceph1'
#对crash的daemons进行归档
[root@ceph1 ~]# ceph crash archive-all
[root@ceph1 ~]# ceph crash ls-new
#之后，集群回复正常
[root@ceph1 ~]# ceph health detail
HEALTH_OK
参考：参考日志
备注
对集群中副本操作常用命令：
noin 通常和noout一起用防止OSD up/down跳来跳去
noout MON在过了300秒(mon_osd_down_out_interval)后自动将down掉的OSD标记为out，一旦out数据就会开始迁移，建议在处理故障期间设置该标记，避免数据迁移。
(故障处理的第一要则设置osd noout 防止数据迁移。ceph osd set noout ,ceph osd unset noout)
noup 通常和nodwon一起用解决OSD up/down跳来跳去
nodown 网络问题可能会影响到Ceph进程之间的心跳，有时候OSD进程还在,却被其他OSD一起举报标记为down,导致不必要的损耗，如果确定OSD进程始终正常
可以设置nodown标记防止OSD被误标记为down.
full 如果集群快要满了，你可以预先将其设置为FULL，注意这个设置会停止写操作。
(有没有效需要实际测试)
pause 这个标记会停止一切客户端的读写，但是集群依旧保持正常运行。
nobackfill
norebalance 这个标记通常和上面的nobackfill和下面的norecover一起设置，在操作集群(挂掉OSD或者整个节点)时， 如果不希望操作过程中数据发生恢复迁移等，可以设置这个标志，记得操作完后unset掉。
norecover 也是在操作磁盘时防止数据发生恢复。
noscrub ceph集群不做osd清理
nodeep-scrub 有时候在集群恢复时，scrub操作会影响到恢复的性能，和上面的noscrub一起设置来停止scrub。一般不建议打开。
notieragent 停止tier引擎查找冷数据并下刷到后端存储。
cpeh osd set {option} 设置所有osd标志
ceph osd unset {option} 接触所有osd标志
使用下面的命令去修复pg和osd
ceph osd repair ：修复一个特定的osd
ceph pg repair 修复一个特定的pg，可能会影响用户的数据，请谨慎使用。
ceph pg scrub：在指定的pg上做处理
ceph deep-scrub:在指定的pg上做深度清理。
ceph osd set pause 搬移机房时可以暂时停止读写，等待客户端读写完毕了，就可以关闭集群
101
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
