[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: How come fsck still kicks in and reports major errors with Ext3?



That's not exactly what I meant.. What I meant was that if you umount
uncleanly in Redhat 8.0, when you reboot it appears runs a FORCED fsck
on the root mount point "/". It takes a while, much longer than a simple
journal replay. The first few times I noticed this I thought I had
accidently formatted "/" as ext2, but it was in fact ext3.

I admit that I don't know why this happens. It may be that each time I
noticed this behavior there really was some error that caused the normal
journal replay to fail. I haven't spent a lot of time testing redhat 8.0
in this area, but that behavior stood out like a sore thumb.

It was as if they had replaced their 7.3 question "filesystems were
umounted uncleanly, press y to run forced fsck" boot question with an
automatic fsck -f.


On Tue, 2002-10-29 at 12:10, Stephen Tweedie wrote:
> Hi,
> 
> On Mon, Oct 28, 2002 at 09:06:02AM -0500, Darrell Michaud <dmichaud wsi com> wrote:
> 
> > It's been my experience with Redhat 8.0 that it is configured by default
> > to run fsck on "/" for every unclean shutdown. Look for the presence of
> > "/.autofsck". Damn annoying.
> 
> That's right --- after an unclean shutdown, fsck will run, but it will
> only perform a journal replay.  fsck on ext3 does *NOT* scan the full
> filesystem unless the forced fsck interval has elapsed or the
> filesystem has been marked as containing an error that needs fixing.
> 
> A journal replay is always required after an unclean shutdown ---
> that's the whole point of the journal.  It doesn't matter if the fsck
> does the replay at startup time, or the kernel does it later on at
> mount time, you still won't get a full fsck unless there is something
> else forcing it.
> 
> Cheers,
>  Stephen

Attachment: signature.asc
Description: This is a digitally signed message part


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]