On Fri, 2004-09-17 at 16:11, Petr Humlicek wrote:
> I'm so ashamed. If I attentively read man, I can see, that I may end information 
> with -v parametr. I can save the time.
> So: I run fsck on test partiton
>                                                                                      
>                                                                             
> [EMAIL PROTECTED] root]# jfs_fsck -fv /dev/hdg6
> jfs_fsck version 1.1.7, 22-Jul-2004
> processing started: 9/17/2004 22.47.48
> The current device is:  /dev/hdg6
> Open(...READ/WRITE EXCLUSIVE...) returned rc = 0
> Primary superblock is valid.
> The type of file system for the device is JFS.
> Block size in bytes:  4096
> Filesystem size in blocks:  29304560
> **Phase 0 - Replay Journal Log
> LOGREDO:  Log already redone!
> logredo returned rc = 0
> **Phase 1 - Check Blocks, Files/Directories, and  Directory Entries
> File system object MF1 has invalid descriptor (C).

Well, I have a bug to fix in jfs_fsck.  At a minimum, this message
should print without the -v flag, and fsck should not return success. 
Better yet, would be to try to repair the problem.

But now, to try to fix your filesystem...

Running "jfs_debugfs /dev/hdg6" should give you a command prompt.  From
here, type "in 1 a".  On a good file system, you should see something
like this:

> xt 1 a
Root X-Tree Node of inode 1

[1] flag        0x83    BT_ROOT  BT_LEAF
[2] nextindex   3               [5] self.addr1  0x00
[3] maxentry    18              [6] self.addr2  0x00000000
[4] self.len    0x000000            self.addr   0
xtree: Hit enter to see entries, [m]odify, or e[x]it:

jfs_fsck is complaining about field [1].  You can fix this by typing "m
1 0x83".  I don't know how much more is corrupt, so this may only get
you as far as the next error, but it's worth trying.

A more bold approach may be to copy the entire block of inodes from a
good jfs filesystem.  You can do this with:
"dd if=/dev/good_filesystem of=/dev/bad_filesystem bs=4096 skip=11
seek=11 count=1"

This should at least get you further, but I don't know how far.

_______________________________________________
Jfs-discussion mailing list
[EMAIL PROTECTED]
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion

Reply via email to