[rfc] use ext3 instead of ext2

Bryan Kadzban bryan at kadzban.is-a-geek.net
Fri Jul 16 03:57:42 PDT 2004


Kevin P. Fleming wrote:
> Archaic wrote:
> 
>> Keep the fsck code in. There is no reason to pull it when a simple
>> 0 0 in fstab can let the user decide for himself. I for one still
>> fsck ext3 systems. Also, readonly is still perfectly valid for
>> bootup.
> 
> Actually I don't think that's true with the current LFS bootscripts;
> I use XFS as my root filesystems here, and even with "0 0" in the
> fstab it stills tries to fsck them.

Yes, but the fsck finishes immediately even in the case of an unclean
shutdown, right?  So what is it actually hurting to have it run?

If it doesn't exit quickly in the case of an unclean umount, then IMO
XFS is hosed (or it needs a fsck tool that knows how to replay its
journal)...

And ISTM that this ONLY applies to the root filesystem, right?  Other
partitions honor the 0 0?  From a cursory look at the bootscripts, they
should.

With my ext3 root FS, in the case of an unclean shutdown, the kernel
replays the journal at the next boot.  Yes, fsck runs later, but it sees
that the FS is clean at that point (because I mount it read-only), and
skips it.  There is the point where during the journal replay it's read
write, but that doesn't seem to affect anything -- fsck still sees it as
being clean.

With ext3 non-root FSes, on an unclean shutdown, they get their journals
replayed by fsck.ext3 on the next boot (in checkfs).



More information about the lfs-dev mailing list