lkml.org 
[lkml]   [1998]   [Sep]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: is this fs corruption?
Hi,

On Fri, 11 Sep 1998 07:16:44 +0200 (MET DST), Ragnar Hojland Espinosa
<ragnar@lightside.ddns.org> said:

> On Thu, 10 Sep 1998, Andrea Arcangeli wrote:
>> On Thu, 10 Sep 1998, Matthias Urlichs wrote:
>> >Andrea Arcangeli:
>> >> >> > Free blocks count wrong for group #7 (358, counted=375).
>> >> I' ve seen this message again after running 2.1.120 but this time 2.1.120
>> >> has not unmounted the fs clean so fsck recovered the problem at the first
>> >OUCH! That hurts...
>> I dubit of hardware problems since 2.1.120 was aware of the problem and
>> was unable to unmount the fs. I am not running overclocked from ages. Here

> I occasionally notice the same problem (free blocks count wrong),
> almost always on the same groups. Is there a way to find out what they
> correspond to?

It is not usually a problem.

In most cases, in the absense of other filesystem errors, finding the
free blocks count wrong will be due to an unclean shutdown while there
were files (typically log files) open for sequential write. This is a
result of a very effective internal ext2fs optimisation.

When you are writing to a sequential file, ext2fs allocates more than
just one block at a time. For performance, and to reduce file
interleaving if you are extending several files at once, it
provisionally allocates a chunk of disk space at once. If you write one
block of data, then only one of these "preallocated" blocks will
actually be allocated to the file; the others remain internally reserved
for potential use by the file until the file is closed. On close, all
of the preallocated blocks are returned to the filesystem's free space.

So, the free block count wrong error is typically just a sign that you
shutdown with such a writable file still open, and the preallocated
blocks were still marked as not free in the fs bitmaps despite not being
used by any file. Because these blocks are genuinely not in use, there
is no way to find out what they correspond to: they are harmless.

--Stephen


-
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/faq.html

\
 
 \ /
  Last update: 2005-03-22 13:44    [W:0.067 / U:0.264 seconds]
©2003-2014 Jasper Spaans. Advertise on this site