pg 3.c is active+clean+inconsistent, acting [18,17,19]

n0bie

Member
Dec 28, 2021
71
2
13
37
i got this error in my ceph status:

Code:
pg 3.c is active+clean+inconsistent, acting [18,17,19]
pg 3.d is active+clean+inconsistent, acting [16,17,19]
pg 3.15 is active+clean+inconsistent, acting [18,19,17]

after some search and investigate.. i notice this:

Code:
root@c1:~# rados list-inconsistent-obj 3.c --format=json-pretty
root@c1:~# rados list-inconsistent-obj 3.d --format=json-pretty
root@c1:~# rados list-inconsistent-obj 3.15 --format=json-pretty
{
                    "osd": 17,
                    "primary": false,
                    "errors": [
                        "read_error"
                    ],
                    "size": 4194304
                },


i suspect osd.17 has a problem in their HDD storage.

check SMART give this information:
S.M.A.R.T : smart failure failure prediction threshold exceeded ascq 0x5

do u think that is safe enough for me to take out disk osd.17 and replace new one?
osd out
osd down
osd destroy
osd create new (with new disk).