On Tue, 3 Jul 2001, Joe Cooper wrote:
>Date: Tue, 03 Jul 2001 13:52:47 -0500
>From: Joe Cooper <joe@xxxxxxxxxxxxx>
>To: kickstart-list@xxxxxxxxxx
>Reply-To: kickstart-list@xxxxxxxxxx
>Content-Type: text/plain; charset=us-ascii; format=flowed
>Subject: Re: Reiserfs
>
>It cannot. Red Hat is not a big fan of ReiserFS, or at least have not
>been to date and don't seem too inclined to support it.
That is not an accurate picture at all. Red Hat believes in
releasing quality products that do not have known data corrupting
issues. We strive to provide solutions that our customers and
users can rely on - which has passed intense internal quality
testing. Reiserfs has not proven itself yet as rock solid in
testing, and even though many now claim it to be bug free,
reports still surface on mailing lists and in testing of data
corruption.
Reiserfs is nonetheless getting better, which is a good thing.
We all would love a good journalled FS, however compromising data
integrity for feature list checkboxes is not good for our
customers.
Many people now use reiserfs and experience no problems, however
many != all, and very large very busy machines with huge
filesystems are the ones most likely to encounter problems,
perhaps problems one would never ever see on a, home machine or
even a small business server.
Keep in mind there is no benefit to excluding something like
reiserfs 'just because'. There are very sound technical reasons
why it is only supported as experimental right now.
>The kernel does have ReiserFS compiled in as a module, so
>ReiserFS can be used once the system has rebooted. Note,
>however, that it is compiled with REISERFS_CHECK enabled, which
>is generally not a good idea for production kernels (not sure
>why Red Hat have done so).
Because we do not feel that the data corruption that many
customers are likely to experience with this experimental support
is something that we can consider "good". By having the checks
in place, those brave enough to experiment with reiserfs that
have problems, will now also have useful debugging data, etc.
with which bugs in reiserfs can be found and fixed which will
result in reiserfs being less buggy sooner, and actually a
supported filesystem, which i believe is what everyone is wanting
no?
>Of course, I'm basing this assertion on the Rawhide kernel,
>rather than the one that ships with 7.1 so maybe I'm wrong.
They both AFAIR enable the checks - to help find and fix bugs.
Linux doesn't just run on small home computers and small
businesses anymore. It runs on big iron also, and quality and
reliability across the board trumps whiz-bang features IMHO.
Hopefully this clears up any bad impressions you may have
misunderstood about reiserfs inclusion.
Personally I believe reiserfs will be ready for widespread usage
by years end if not sooner. Lots more people are using it now so
bugs are found faster.
Take care,
TTYL
----------------------------------------------------------------------
Mike A. Harris - Linux advocate - Open Source advocate
Opinions and viewpoints expressed are solely my own.
----------------------------------------------------------------------
The day Microsoft makes something that doesn't suck, is
probably the day Microsoft starts making vacuum cleaners.
