Lindsay Haisley <fmouse-gentoo@...> posted
20070206181629.GA19923@..., excerpted below, on Tue, 06 Feb 2007
> Thus spake Daniel Gryniewicz on Tue, Feb 06, 2007 at 10:19:17AM CST
>> In general, reiserfs is considered dead by the linux kernel guys, and by
>> it's last remaining maintainer in particular
> Why do I always seem to pick loosers when it comes to software technology.
> All the important partitions on two of my servers and on my desktop
> Gentoo box are reiserfs. I suppose in a couple of years kernel support
> for reiserfs will start to get flakey and I'll be in a pickle unless I
> rebuild the boxes.
It'll be more than a couple years, as there's a LOT of installed base out
there for it. Remember, SuSE and others defaulted to reiserfs for many
years, and people are running servers on it and won't want to change until
they die and are replaced, if they can help it. Since it's those guys
that pay the money to the distributions that in turn pay many of the
kernel devs, it's a safe bet to say reiserfs support will be around for
a good while -- almost CERTAINLY for another hardware generation or two.
Also, keep in mind that even in the linked post, he was talking about
globalfs as the replacement, and he said it wasn't ready and wasn't
/going/ to be ready for awhile.
IMO, reiserfs will continue to be a solid choice for several more years
anyway, during which time filesystems technology will continue to
evolve. Thus, at least for those comfortable with reiserfs now, there's
no need to get worried for another hardware upgrade or two, by which
time the choices will have likely evolved to the point where an upgrade
path is reasonably clear, and those switching now just to get off the
supposedly dying reiserfs, will be kicking themselves for not waiting, and
thus either having to do yet another fs migration, or watching everybody
else play with the new best choice while they are stuck with what was the
best choice back in early 2007.
Think about it. Say three years ago, who would have thought xfree86 would
be past history for most Linux users by now? Who would have predicted
xorg. Who would have predicted the drastically different kernel 2.6
development model, with Andrew's -mm becoming what amounts to the
developer kernel, and no 2.7 in sight? On the other side, who would have
predicted Hans Reiser would get into the legal trouble he's seeing now,
accused of killing his wife?
We see what happened in the last three years. How can we possibly look
ahead three years and predict with any certainty what sorts of other
drastic Linux world altering events will happen, and what that will do to
shape the choices we have at that time in the mean time? While I'd not
necessarily recommend reiserfs for newbies to get into, those using it and
already comfortable with it have no reason to change right away if they
are happy with it, since there's nothing close to matching its
particular featureset, and every reason to wait awhile, thus being better
positioned to examine the opportunities as they arise, without having to
jump filesystems twice in three to four years!
If you are switching hardware and have a good reason to change ATM, do so.
If you are comfortable where you are, there's every reason to believe
that's not going to change significantly thru at least this hardware
upgrade cycle. At the next one, examine where things are, and go from
That's what I'm doing, anyway, because I happen to believe it's the most
solid and logical choice available, under the circumstances.
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman
email@example.com mailing list