On Sun, Sep 22, 2019 at 9:25 PM Arkadiusz Miśkiewicz <a.miskiewicz@xxxxxxxxx> wrote: > > On 16/09/2019 23:35, Eric Sandeen wrote: > > On 9/15/19 7:44 AM, Arkadiusz Miśkiewicz wrote: > >> > >> Hello. > >> > >> xfsprogs 5.2.1 and: > >> > >> disconnected dir inode 9185193405, moving to lost+found > >> disconnected dir inode 9185193417, moving to lost+found > >> disconnected dir inode 9185194001, moving to lost+found > >> disconnected dir inode 9185194004, moving to lost+found > >> disconnected dir inode 9185194010, moving to lost+found > >> disconnected dir inode 9185194012, moving to lost+found > >> disconnected dir inode 9185194018, moving to lost+found > >> disconnected dir inode 9185194027, moving to lost+found > >> disconnected dir inode 9185205370, moving to lost+found > >> disconnected dir inode 9185209007, moving to lost+found > >> corrupt dinode 9185209007, (btree extents). > >> Metadata corruption detected at 0x449621, inode 0x2237b2aaf > >> libxfs_iread_extents > >> xfs_repair: phase6.c:1129: mv_orphanage: Assertion `err == 2' failed. > >> Aborted > > > >> > >> > >> # grep -A1 -B1 9185209007 log > >> entry ".." at block 0 offset 80 in directory inode 9185141346 references > >> non-existent inode 6454491396 > >> entry ".." at block 0 offset 80 in directory inode 9185209007 references > >> free inode 62881485764 > >> entry ".." at block 0 offset 80 in directory inode 9185220220 references > >> free inode 6454492606 > >> -- > >> rebuilding directory inode 9185141346 > >> entry ".." in directory inode 9185209007 points to free inode > >> 62881485764, marking entry to be junked > >> rebuilding directory inode 9185209007 > >> name create failed in ino 9185209007 (117), filesystem may be out of space > > > > 117 is EUCLEAN/EFSCORRUPTED even though we were in the process of rebuilding it. :( > > > > so this is probably why a subsequent attempt to move it to lost+found failed as well? > > > > Is this a metadumpable filesystem...? > > > It's one of my big fses but metadump can't deal with it: > > > [...] > > Copied 103433024 of 165039360 inodes (24 of 39 AGs) Unknown directory buffer type! > > Copied 104001280 of 165039360 inodes (24 of 39 AGs) Unknown directory buffer type! > > Copied 105465088 of 165039360 inodes (24 of 39 AGs) Unknown directory buffer type! > > Copied 107092608 of 165039360 inodes (25 of 39 AGs) Metadata corruption detected at 0x473455, xfs_dir3_leaf1 block 0xc8471ce78/0x1000 > > Segmentation fault (core dumped) You could still try to dump without data zeroing (metadump -a), if this is acceptable to you.
