Re: [ext2resize] ext3, lilo?

[Andreas D. <ad...@cl...>]

On Apr 13, 2003  01:51 -0400, ei...@th... wrote:
> I'm writing up documentation for "how to turn a Martian NetDrive into
> an OpenAFS server", and one of the steps is, simply,
> 
>  boot to single user mode
>  ext2resize / 4G
>  reboot
> 
> Now, this *worked* for me, but on looking at some of the docs, I have
> a couple of questions.
> 
> 1) / is actually an ext3 partion, not ext2.  However, having been
>    unmounted cleanly, the journal should be empty, so it should be
>    identical in that regard.  Is this actually true, or did I just get
>    lucky, and should I tell people to delete the journal, remount as
>    ext2, resize, then tune2fs -j and make it ext3 again?

This is OK to do on an ext3 filesystem.  As you say, the journal is empty
after a clean shutdown/unmount, so there is no problem with this.  If it
wasn't cleanly shut down, there would be an extra flag "needs_recovery"
that signals data in the journal and ext2resize would refuse to resize.

> 2) I didn't rerun lilo.  I suspect that since the partition in
>    question had around 1G of data on it, that there was simply nothing
>    beyond the cut-off point, so nothing had to be moved.  

Correct.

>    a) is that true, that in that case nothing would move?

Yes.

>    b) would -v be enough to find out if it moved something, namely the kernel?
>    c) should I tell people to re-run lilo anyway?

Couldn't hurt, but unnecessary.

Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/