Possible data damage 1 pg recovery_unfound

参考：
https://ceph.com/geen-categorie/ceph-manually-repair-object/

查看ceph状态

[root@node-1 ~]# ceph health detail
HEALTH_ERR 1 scrub errors; Possible data damage: 1 pg inconsistent
OSD_SCRUB_ERRORS 1 scrub errors
PG_DAMAGED Possible data damage: 1 pg inconsistent
    pg 2.9b is active+clean+scrubbing+deep+inconsistent+repair, acting [38,41,13]

1.修复pg
一般情况采用 ceph pg [pgid]，但是经过观察，并不能解决。

[root@node-1 ~]# ceph pg repair 2.9b

2.查询pg使用的主机osd

[root@node-1 ~]# ceph pg 2.9b query|grep primary
            "same_primary_since": 29426,
                "num_objects_missing_on_primary": 0,
            "up_primary": 38,
            "acting_primary": 38,
                "same_primary_since": 29426,
                    "num_objects_missing_on_primary": 0,
                "up_primary": 38,
                "acting_primary": 38,
                "same_primary_since": 29426,
                    "num_objects_missing_on_primary": 0,
                "up_primary": 38,
                "acting_primary": 38,

3.查找osd对应的物理节点host

[root@node-1 ~]# ceph osd find 38
{
    "osd": 38,
    "addrs": {
        "addrvec": [
            {
                "type": "v2",
                "addr": "10.1.2.10:6816",
                "nonce": 3635776
            },
            {
                "type": "v1",
                "addr": "10.1.2.10:6817",
                "nonce": 3635776
            }
        ]
    },
    "osd_fsid": "1d3dd4ee-1303-4709-9fa0-e8120a930d0f",
    "host": "node-10",
    "crush_location": {
        "host": "node-10",
        "root": "default"
    }
}

4.停掉osd

[root@node-10 ~]# systemctl stop ceph-osd@38.service

5.刷入日志（echo $?,如果不等于0先不用管，尝试启动osd看一下）

[root@node-10 ~]# ceph-osd -i 38 --flush-journal

6.启动osd

[root@node-10 ~]# systemctl start ceph-osd@38.service

7.查看pg所在osd

[root@node-1 ~]# ceph health detail
HEALTH_OK

[root@node-1 ~]# ceph pg ls|grep 2.9b
2.9b     4164        0         0       0 15635105244           0          0 3005           active+clean   78m 29447'25665617 29447:26608464 [38,41,13]p38 [38,41,13]p38 2022-08-08 11:32:30.636330 2022-08-08 11:32:30.636330

可以确认集群恢复OK。PG还是在 osd.38上。