I suppose I could check myself, but is it possible to print the block # in
which the bad sector was discovered and in turn which inode that block
belongs to and in theory the filename of the file that just got munged?
That'd have to be in the ext2 fs I suppose.
It'd be ever so nice to know what file just bit the dust so I could know
what to replace other than say, compairing all the files against a tape
backup that's too old to really be of any use.
> > Jul 5 15:26:31 soda kernel: hda: dma_intr: status=0x51 { DriveReady
> > SeekComplete Error }
> > Jul 5 15:26:31 soda kernel: hda: dma_intr: error=0x40 {
> > UncorrectableError }, LBAsect=626563, sector=9666
> > Jul 5 15:26:31 soda kernel: end_request: I/O error, dev 03:02, sector
> > 9666
BTW, I personally don't think that the error messages are that mysterious,
and I don't see why people immediately think ext2 errors, since no part of
the error messages mention ext2. Ext2 actually remains awfully silent as
your HD blows chunks. Perhaps explicitly declare it as an IDE error, not fs
error. Such as:
Jul 5 15:26:31 soda kernel: IDE (hda2) drive DMA error: Error occured after
Drive is Ready and Seek Completed (status=0x51).
Jul 5 15:26:31 soda kernel: IDE (hda2) drive DMA error: Unable to Correct
Error, LBAsect=626563, sector=9666
Jul 5 15:26:31 soda kernel: IDE (hda2) Possible drive failure - Check
cabling and connections.
> Maybe we should have the IDE driver print something about "likely drive
> failure", and also more clearly identify which drive is failing. "dev
> 03:02" is simply not very descriptive for most poeple.
Printing the major/minor # in decimal would at least help the hexadecimally
challenged anyway.
- Steve
.------------------------------------------------. # * # # # # # #
| Steve Baker | Barely Working | # ## # # # # #
| ice@mama.indstate.edu | System Administrator | # # # # # # # #
| Red-Hat Rulz! | Will work for hardware | # # # ## # # # #
`-- SYS-ADMIN FOR HIRE, HAVE UNIX, WILL TRAVEL --' #### # # # ## # #