It's an on disk corruption, and hence can be fixed (if at all) only by fsck. Question is why are you running into these corruptions so frequently. Check if there are any issues with the disk/storage or any applications accessing the disk directly. Also, you are running very very old version of the file system, please upgrade and see if that helps.
On 5/25/2011 11:41 PM, Alex Sobrino wrote: > El 20/05/11 10:03, Srinivas Eeda escribió: >> That inode number looks suspicious. Appears fs might have got >> corrupted, can you umount the filesystem and run fsck.ocfs2 -f to >> check for any corruptions. > > Hi, > > Last weekend I umounted the filesystem and permorfed and fsck.ocfs2 > -f. It found a lot of errors that seemed to be fixed. In fact, after > the fsck.ocfs2 I could create/delete files from the /2712/ directory. > In fact, I restored it's last backup. > > Since yesterday, we have (again) the same problem in the same > directory (I can't be sure if we also have it in another directory). > When I try to perform any I/O operation into that directory, the logs > says: > > May 26 08:32:05 swmdl1 kernel: (21444,0):ocfs2_check_dir_entry:155 > ERROR: bad entry in directory #58127027: directory entry across blocks > - offset=0, inode=3709498566882687729, rec_len=65504, name_len=0 > > And all the files in that directory are missing. Is there anything I > could do besides a fsck.ocfs2? I know that could solve the problem, > but I can't be sure if it'll come back again :S > > Thanks for all, > _______________________________________________ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
