Where is the script that does the automatic e2fsck's located?

[Robert Locke]

On Fri, 2005-03-18 at 01:29 -0500, Gene Heskett wrote:
> On Thursday 17 March 2005 22:57, Mathieu Chouquet-Stringer wrote:
> >paul at city-fan.org (Paul Howarth) writes:
> >> > are you guys gonna keep batting this around?
> >> > man fstab
> >> > see explanation on 6th field
> >>
> >> What has that got to do with forcing a filesystem check?
> >
> >Nothing but that wasn't the original question.
> >
> >> Gene doesn't want to disable the filesystem check, he just wants
> >> it to be more verbose so it's clear that the system hasn't hung.
> >
> >Quote:
> >"Filesystems within a drive will be checked sequentially, but
> > filesystems on different drives will be  checked  at  the  same
> > time  to  utilize parallelism  available  in the hardware."
> >
> >Meaning that (using Gene's config with hda and hdd), if you happen
> > to have hdaX and hddY and that fsck starts to check them at the
> > same time (or hdaX first), you won't see progress for hddY.
> 
> I'm not seeing any progress for /dev/hdd3, ever.  Having another 
> running at the same time on a different spindle has nothing to do 
> with the problem I'm reporting.
> 
<snip>
Actually, Gene, that's exactly the point.  Only one of the perhaps many
parallel running fsck's is displaying it's output on your console.  In
your case, with two spindles, you will have the partitions on the
"second drive" effectively being checked in the background because the
partitions on the "first drive" are taking console ownership if you
will.  So while hda1 and hdd1 are both being checked, you would see only
hda1, and then when hdd2 starts, well, hda2 already has the console, and
so on....

One option to prevent the parallelism, and cause your system to take
even more time to boot, might be to adjust the final column
in /etc/fstab.  If you change the 2's on the hdd partitions to 3's,
then, if memory serves, hdd would be checked in a 3rd pass, after pass 2
has finished hda, and you should then see the output from checking hdd
since we are not then running in parallel, but rather sequentially.

HTH,

--Rob