After full recovery as described here
https://docs.ceph.com/docs/master/cephfs/disaster-recovery-experts/#disaster-recovery-experts
there are still problems in filesystem.
"loaded dup inode" error log messages lead me to /lost+found folder
loaded dup inode 0x10000ed6c1a [head,head] v1 at /lost+found/10000ed6c1a, but inode 0x10000ed6c1a.head v493 already exists at ...
I tried to remove files in /lost+found and got error "Read-only file system"
rm: cannot remove '10000ed02dc': Read-only file system
scrub start / repair failed with
{
"performed_validation": true,
"passed_validation": false,
"backtrace": {
"checked": true,
"passed": true,
"read_ret_val": 0,
"ondisk_value": "(2)0x1:[]//",
"memoryvalue": "(2)0x1:[]//",
"error_str": ""
},
"raw_stats": {
"checked": true,
"passed": false,
"read_ret_val": 0,
"ondisk_value.dirstat": "f(v0 m2020-05-31T06:04:03.924797+0000 2=0+2)",
"ondisk_value.rstat": "n(v0 rc2020-06-01T12:00:00.000000+0000 b397013380709 5466639=3777584+1689055)",
"memory_value.dirrstat": "f(v0 m2020-05-31T06:04:03.924797+0000 2=0+2)",
"memory_value.rstat": "n(v19075 rc2020-06-01T12:00:00.000000+0000 b397013380709 5466639=3777584+1689055)",
"error_str": "dirfrag(*) has bad stats; "
},
"return_code": 0
}
scrub start /lost+found repair didn't helped too
{
"performed_validation": true,
"passed_validation": false,
"backtrace": {
"checked": true,
"passed": true,
"read_ret_val": 0,
"ondisk_value": "(2)0x4:[<0x1/lost+found v1>]//",
"memoryvalue": "(2)0x4:[<0x1/lost+found v454066>]//",
"error_str": ""
},
"raw_stats": {
"checked": true,
"passed": false,
"read_ret_val": 0,
"ondisk_value.dirstat": "f(v0 1=1+0)",
"ondisk_value.rstat": "n(v0 rc2020-05-31T06:07:50.359058+0000 1=0+1)",
"memory_value.dirrstat": "f(v1 1=1+0)",
"memory_value.rstat": "n(v1 rc2020-05-31T06:07:50.359058+0000 1=0+1)",
"error_str": "dirfrag(*) has bad stats; "
},
"return_code": 0
}
How can I repair this?
