[linux-help] Re: HD trouble

[james l <james@xxxxxxxxxxxxxxxxxx>]

> I've been getting good info from this mailing list and 
> making great strides toward setting up my linux 
> ftp/mail server.  I started by buying a new 30GB drive 
> for my little 486DX.  It was working great until a day 
> ago.  At first I heard a high-pitched whistling sound 
> from the computer.  Then, about a day later, I started 
> getting "ext2-fs error" messages.  Now I'm getting 
> them all the time.  During boot this last time, it 
> gave me the option to enter "maintenance mode".  I ran 
> fsck.  It determined I had "2.6% non-contagious 
> errors."  I'm looking for a way to get more info on 
> what's going on.  Is fsck the only tool I've got, or 
> is there something else I can use as well?
> 
For the whistling: I have no idea.

If you crashed the machine (power off w/o halting, kenel panic, etc), the file
system (ext2) can become have errors. I have occasionally had to enter
maintenance mode, due to this, usually accepting what e2fsck says to do works
(has for me anyway), but if it really trashed the filesystem, well....bad
luck. Normally though, it may have only a small number of unimportant errors
(eg inode xxxx has deltime of zero-FIXED) which is handled if it wasn't
umounted properly. It will display how contagious the data is, or how many
files are not in one 'steam' from the disk, because they are stored in
different locations on the disk via pointers. This is also known as
fragmentation, but ext2 doesn't have much. It also depends on the number of
read/writes of files on the partition (eg, /usr usually has <1% while a
particularly active /home may have 5% or occasionally more).

The whistling may indicate that something is wrong with the drive, or the
warnings may indicate bad blocks. Linux has a tool to check for them called
badblocks. Running "badblocks /dev/hda8" (replacing hda8 with what yours is)
should tell you if there are really bad blocks (make sure it is umounted, or
mounted read-only) If there are you can output which ones are bad to a file,
and have e2fsck mark them as bad for the file system, or you might check your
warranty. (see man pages  badblocks(8) and e2fsck(8) for more info)

Hope this helps, and i hope I am right about everything.
James L

-- This is the linux-help@xxxxxxxxx list.  To unsubscribe,
visit http://tmp2.complete.org/cgi-bin/listargate-aclug.cgi