> > > Net Llama! <[EMAIL PROTECTED]> wrote:

> > >>>>Just to stire the pot, and play devil's advocate, i will say
> > >that>>>it hasn't achiveved respectability in alot of groups.  There
> > >are>>>some heated attacks on gentoo on the XFS mailign list, and
> > >i've>>>seen some snide remarks on LKML as well.
> > >>>>
> > >>>

> > >>
> > >>The majority of the snide remarks come from people who have looked
> > >at>Gentoo's special kernel patches and been revulsed.  There's a
> > >>disturbing corellation between people reporting unreproducable XFS
> > >>filesystem corruption and using Gentoo's kernels.
> > >>

I did a little research on both the XFS comments about gentoo and on the
gentoo-user list. Here's my take on the situation:

1. Gentoo offers all sorts of kernels, from plain vanilla to every sort
of experimental kernel imaginable. There aren't any current statistics,
but, if last year was any indication, a lot of people go for the
experimental kernels. For me, the vanilla kernels are good enough.

2. The most recent XFS complaints I could find about gentoo were a
mixture of 2.6 32-bit and 64-bit users, and yes most of these were
experimental kernels. There are only a handful of reports from gentoo
users.

3. The XFS developers choose to ignore any reports using an experimental
kernel. I'm of two minds about that approach. 1) Sounds fine. It's more
difficult to analyze with extra patches. 2) The reports aren't "my
kernel broke" but "my file system crapped out." If the XFS code is
really fragile enough that the presence of some extra patches is a
problem, then maybe some extra work needs to be done by the developers.
I suspect that some of the reported problems are indeed XFS bugs that
will go unsolved because the developers don't like to deal with extra
patches. I haven't seen a lot of reports from ext3 users on experimental
kernels saying "my file system crapped out.:- reiserfs yes, even on
vanilla kernels.

4. The comments in the XFS postings complaining about gentoo are about
as non-specific as you can get. "They include patches they shouldn't and
they omit patches they should have." "Get an RH asap." It would be nice
to have some real examples and some analysis of what the presence /
absence of patches has to do with the filesystem layer. Where's the
beef, i.e. the proof that the problem has anything to do with extra
patches? Not all kernel patches affect the filesystem layer. There's a
similar situation with "I don't like the version of glibc/gcc your're
running." As above, no real information, no taking the opportunity to
do advance work to deal with the glibc/gcc changes which will just come
back to bite later with even more users, just griping.

5. I've ignored the comments about 2.4 kernels. In those days XFS was
patch city. It should be a little easier for everyone now that
XFS is in the kernel source tree.

6. The hard feelings go both ways. The gentoo developers were burned
badly by an XFS server they were running a while ago, and the XFS
developers hate to see problem reports from gentoo because of the
experimental kernels. In an ideal world, there would be more cooperation
and less animosity, but unfortunately the trend is toward animosity.
Neither attitude is to be commended.

In summary, gentoo is respected by a growing number of users, but
only tolerated by a large number of others. End of story for my part.

Enjoy,

-- 
 /\/\
( CR ) Collins Richey
 \/\/     fly Independence Air - they run Linux



_______________________________________________
[EMAIL PROTECTED]
Unsub/Pause/Etc -&gt; http://mail.linux-sxs.org/cgi-bin/mailman/listinfo/general

Reply via email to