Re: 2.1.x on UX -- How?

[Adrian Gartland <ade oregan net>]

On Tue, 17 Nov 1998, furio ercolessi wrote:

> > Easiest fix is to use the latest MILO and ldmilo.exe for UX/RUFFIAN, found
> > on gatekeeper.dec.com as:
> > 
> >    /pub/DEC/Linux-Alpha/Kernels/ldmilo-ruffian-981020
> >    /pub/DEC/Linux-Alpha/Kernels/milo-ruffian-981020

We upgraded to these a few days ago..before we considered
upgrading to 2.1.x....so yes...it was with this that it was failing.
Going back to the original milo resulted in exactly the same.
It wasn't me that actually did it..so guess I will do so to make
sure.

The kernel was only 100k bigger than the 2.0.35..does it make
that much difference?
 
> Just out of curiosity (we switched to LX's one month ago, and
> we are very happy with them :-) : is the "Unable to handle paging
> request" problem now solved with this Ruffian MILO?
> And if yes, where was the problem?

Havn't actaully seen one of these for a while...odd.
Been getting a rather large amount of
Whoops: unlock_buffer: async io complete on unlocked page
lately tho :(...putting the process into an uninterruptable
sleep permanently. Most annoying. At least that doesn't toast 
the machine.

We tried the X server out which allows >25 xterms....
There hasn't been a single problem with that machine
since it was done (1.5 days)...hmmm.

Just out of interest....
Has anyone else noticed pipes or unix domain sockets
being extremely slow compared to i386?
I'm trying out 2.1.x to see if its any better.

Is there a recommended 2.1.x kernel for alpha? Can't
afford to wait till christmas for 2.2 to arrive.

Tar.
-- 
Adrian Gartland              [Email:         ade@oregan.net]
Software Engineer            [WWW:   http://www.oregan.net/]
Oregan Networks              [Phone:     +44 (0)1530 563311]