Messages in this thread | | | Date | Fri, 29 Oct 1999 22:37:25 +0200 (CEST) | From | Urban Widmark <> | Subject | ext2 fs corruption in 2.2.12 |
| |
Hello
I don't know if this means anything to anyone, and since there is a 2.2.13 maybe this is old.
This is what I found in the logs this afternoon:
Oct 29 15:58:33 cola kernel: EXT2-fs error (device ide0(3,6)): ext2_readdir: bad entry in directory #46370: rec_len is smaller than minimal - offset=0, inode=0, rec_len=0, name_len=0 Oct 29 15:58:40 cola kernel: EXT2-fs warning (device ide0(3,6)): ext2_free_blocks: bit already cleared for block 188420 Oct 29 15:59:06 cola kernel: EXT2-fs warning (device ide0(3,6)): ext2_free_blocks: bit already cleared for block 189301 Oct 29 15:59:06 cola kernel: EXT2-fs warning (device ide0(3,6)): ext2_free_blocks: bit already cleared for block 192500 ... Oct 29 16:13:12 cola kernel: EXT2-fs error (device ide0(3,6)): ext2_add_entry: bad entry in directory #50457: rec_len is smaller than minimal - offset=2048, inode=0, rec_len=0, name_len=0 Oct 29 16:17:28 cola kernel: EXT2-fs error (device ide0(3,6)): ext2_add_entry: bad entry in directory #50457: rec_len is smaller than minimal - offset=2048, inode=0, rec_len=0, name_len=0 Oct 29 16:19:24 cola kernel: EXT2-fs warning (device ide0(3,6)): ext2_free_blocks: bit already cleared for block 188423 ... Oct 29 17:40:50 cola kernel: EXT2-fs warning (device ide0(3,6)): ext2_free_blocks: bit already cleared for block 192895 Oct 29 17:40:50 cola last message repeated 2 times ... Oct 29 18:20:02 cola kernel: EXT2-fs error (device ide0(3,6)): ext2_readdir: bad entry in directory #50457: rec_len is smaller than minimal - offset=0, inode=0, rec_len=0, name_len=0
and as you can see it took me quite a while to notice there was something wrong.
I don't know exactly what I did at 15:58:33. The only thing that was special today was that the ibm-jdk-1.1.8 coredumped and filled the partition in question (and that may have been around that time), making procmail put the mail in the inbox instead. A few of my mailfolders were partially filled with '\0', but I first thought that had something to do with procmail & the disk filling up.
e2fsck was not happy with the fs, and it took several runs to get it to be quiet. Some files were lost in the process (but nothing valuable that wasn't on last nights backup). lost+found contains a lot of files with strange permissions. I'm keeping the partition as it is now for a while in case someone has ideas on things too look for.
One last thing; the inodes e2fsck was complaining about were mostly in series. Maybe that's a clue.
Linux 2.2.12, RH6.0, uptime ~10 days, egcs-1.1.2 compiled kernel, e2fsck 1.14
/Urban
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |