lkml.org 
[lkml]   [1999]   [Jun]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: 2.2.x instabilities
hi folks- troublemaker again...

Gregory Maxwell wrote:

> This is not in line with my expirences, I have quite a few 2.2
> boxes under extreme load.

OK, but we might be doing different kinds of things with our systems,
so your mileage may vary.

> Doctor doctor it hurts when I move my arm like this.. Then dont
> move your arm like that!

Lets not be silly, copying files to an NFS filesystem is not an
unnatural act. Anyone following this thread is aware of the serious NFS
problems reported every day (including today). Its difficult to submit
a detailed bug report on 2.2.9-ac1 NFS when the system crashes so bad
that it reboots 5 seconds later. The Linux users I work with have
dumped knfsd entirely and reverted back to user-space NFS from RedHat
5.2. It works and is stable (and faster too).

> > * swap issues.
>
> Care to be more specific?

There is a thread on this subject running right now.... I'm speaking
for Linux users in general, not just for myself.
> > * networking instabilities.
>
> Again? some details?

same as above.

> > * NCP (novell) clients are iffy. I can break it in minutes.
>
> Wouldn't know about that.

That's ok, I don't like NCP either ;( but its what my employer uses, so
I have NO choice except to use NT ( nasty, yuck... )

> > * SMP is still 'delicate'.
>
> Come on! Lets have some details,

Again, several threads in progress.

> > * SMP/Alpha threads/mutexes are unstable.
>
> Wouldn't know, no SMP alpha boxes here.. :(

Not everybody uses intel; there are lots of alpha/sparc/mips/m68k users
(like me) who must wade through oversights introduced by x86-only
developers. (In all fairness, they really have no way to test their
changes on the other arch's, so coordination with users of the other
arch's is important...).

> > * kernel sources often generate errors/warnings. We can do a LOT
> > better here.
>
> Errors? Can't say I've seen that.. As for warnings, the stupid
> compiler doesn't always know best.

True, but smart humans can always eliminate warnings from questionable
code. warnings ARE errors in when it comes to released code.

> > * kernel crashes/oops's. I NEVER got one of these with 2.0.x
> > kernels.
>
> Can you please post your decoded oppses?

I'd love to, but the machine reset itself afterwards.

> > I'll continue to use, promote, and develop <Linux> for the
> > foreseeable future...
>
> Thats good, but it would be much more useful if you would kindly
> produce some useful bug reports (please excuse me if you already
> have and I missed your posts).

Many of my problems mirror those reported on this thread. I prefer to
stay quite unless I have something new to add to the already heavy
traffic on this list, however seeing resources diverted from 2.2.x to
2.3.x before the former is stable is a big mistake. Even though
important fixes will be made to 2.3.x, back-porting to 2.2.x is doubling
the amount of work (split source tree), and a big hassle that is prone
to mistakes.

I've received many personal emails from others on this list echoing
frustration with 2.2.x, who have given up reporting problems due to lack
of response. I'll continue to help all those I can and I know you will
too.

I raise these concerns -only- on this list, which represent a pseudo
private, selective group. The Linux community is the best bunch of
developers on the planet, but we could do just a little bit better (me
included)- rigorously testing our stuff before releasing it to an
impressionable world.

ron

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:52    [from the cache]
©2003-2011 Jasper Spaans