> 1.Is ReiserFS part of the kernel or is it some type of add on? It is
> part of the kernel in that the kernel must understand
> filesystems. You
> can download the code for XFS from SGI. I don't know where to get
> ReiserFS except SuSE. There may be a project site on the web.
http://devlinux.com/projects/reiserfs/
SuSE is the primary sponsor of reiserfs and SuSE includes it
in their 2.2-based distros.
> Filesystems are not a modules due to there nature, but I know
> that only
> a portion of XFS goes into the kernel. XFS has daemons that work with
> the kernel to manage writes and caching of data.
Filesystems can be modules if they aren't necessary for booting.
> 2.Who is this different for the journaling that's to be offered in the
> 2.4 kernel? Well, I am not familiar with what is to be in
> 2.4, however,
> I know that a journaled filesystem is necessary for any large
> filesystem
> to be recovered quickly. Certain types of business and
> projects require
> very short recovery times and fsck'ing a Terabyte of data ain't gonna
> work! I think it is up to us, the Linux community to vote on what we
> want. I like ReiserFS so far, but want to try XFS under
> Linux as it is
> extremely robust under SGI IRIX. The code is out there, we
> just have to try it out.
>
> 3.I would take it you have to install with ReiserFS, you can't take an
> exsisting system and convert it? Actually, if you can back up your
> filesystem, then you can remake it with the XFS or ReiserFS
> tools. Just
> recover the filesystem off the tape and your good to go.
> You'll need to
> understand how to modify your /etc/fstab file so mount knows what type
> filesystem you have. At install time on SuSE 6.4, you get a choice of
> what you want the system to make. I haven't had time to
> investigate the
> ReiserFS tools, but if you download XFS off SGI's website, then you
> would use mkfs_xfs to build the filesystem. Under IRIX you can choose
> the number of inodes, location of the journal/log (put it on
> a separate
> controller for performance with databases), and block size.
> Usual kind
> of filesystem stuff!
>
> I wouldn't feel overwhelmed by all the automated stuff that goes on
> during an install of a Linux distro. Actually, studying that
> process is
> very good to understand recovery procedures since when things
> crap out,
> you sometimes have to go thru most of those steps to get
> things back in
> shape!
>
>
> --
There's also ext3 but I hear that it won't be ready for
2.4.0. Maybe later in the 2.4 lifetime.
IBM JFS is also available:
http://oss.software.ibm.com/developerworks/opensource/jfs/
~Randy
--
To unsubscribe: mail ">majordomo@ale.org with "unsubscribe ale" in message body.