Re: Ext3fs and fsck

[Alan Shutko <ats@acm.org>]

Anthony Campbell <a.campbell@doctors.org.uk> writes:

> It's precisely the above warnings that make me rather nervous of using
> ext3

They shouldn't.  Those warnings apply just as much to ext2, or xfs,
even reiserfs[1].  All it's saying is that unrelated kernel bugs,
drive failures, or other problems may occasionally cause errors, and
you should let fsck run every so often even if you haven't seen
errors, so that it can clean things up before they become big
problems.

This isn't really new... forced fscks after some number of mounts have
been a part of Unix-like operating systems since Unix came out.  It's
just now with ext3, some people want to get rid of all fscks and the
man page has to explain forcefully that that's a bad idea, no matter
how stable the filesystem code is, because things other than the
filesystem code and crashes can cause problems.
> 
> If I leave /etc/fstab as it is, will all the file systems continue to be
> checked as normal when I reboot?

It should.  On my sid system, at least, I still see fsck run and do
forced checks every so many mounts, with no changes to fstab other
than changing ext2 to ext3.  I don't know why the original poster
wasn't seeing things checked... maybe his fstab has 0 in the last
column (ie, telling fsck not to check the fs)?

When adding a journal, tune2fs automatically sets the stuff for forced
fscks every so many mounts or so many months, so unless you overrode
it with -c or -i, you should be fine.

> What is the purpose of the /forcefsck file that is suggested, and what
> commands is it supposed to contain?

If the file exists, an fsck will be forced regardless of other
parameters like fs clean bit, or mount count.  I have never actually
used it, I think the file just needs to be there.

I'm not sure precisely why it's there, but it is.

-- 
Alan Shutko <ats@acm.org> - In a variety of flavors!
"I'd love to go out with you, but I never go out on days that end in `Y.'"