[ILUG] odd nfs problem...

[kevin lyda]

On Fri, Jan 18, 2002 at 11:16:51PM +0000, Paul Jakma wrote:
> On Fri, 18 Jan 2002, kevin lyda wrote:
> > i'm not downgrading to 2.4.  if anything i want to upgrade to
> > 3.0. 
> never mind that, the real question is why you had to have OpenBSD:
> "security dude.. OpenBSD is leet!" were your words iirc¹, and then
> decided to involve nfs.

that's the second dangling footnote this week.  is this some attempt at
corrupting our minds?

and no, openbsd for the router was so that i could set it up and forget
about it.  the fact is that openbsd has a track record of four years with
no remote exploits in it's default config (which i've shaved down even
further).  the fact is that an openbsd remote exploit will be such huge
news when it happens i can pretty much skip monitoring the mailing lists.

the openbsd box i'm describing here is a sparc5.  and i've found linux
on a sparc to, well, suck to be honest.  and i've tried a lot of kernels
over the years.  the fact is that sparc64 and higher end sparcs get
all the attention which is fine by me since *bsd's support older sparcs
quite well.  all i need running is X, everything else can be fired off
other machines, and i find *bsd rather comfy since it's what i started on.

and while you're in a grumpy mood i'll answer the next question: why am
i using a sparc?  because it has the best graphics of any machine in the
house.  i define good graphics as big and can scroll my xterms quickly.
i understand other people define graphics performance in other ways,
good for them.

of course we could have completely avoided this conversation if liam
had paid rent that month with an apple cinema 22" display, g4 cube and
a copy of debian/ppc.  8^}  heck i'd have thrown in breakfast in bed
every morning...

> before this patch my diskless indy was agonisingly slow, constant NFS
> timeouts - server at 100MBit, indy @ 10MBit, cheap hub between them
> whose idea of bridging packets is 'drop them'. after applying this
> patch to the server it worked fine.

could be a theory except that i see 0 bad or dropped packets on the
ethernet layer.  both the server and workstation use 10mb cards and the
linux workstation with no problem is at the same point topologically.

i suspect it's an openbsd error as it is a rather ancient version
of openbsd.

kevin

-- 
kevin at suberic.net          buffy: come on, can't you put your foot down?!
fork()'ed on 37058400      giles: it *is* down.
meatspace place: orbit     buffy: one of these days you're going to have to
http://suberic.net/~kevin         get a grown up car.  --inca mummy girl