dyer@spdcc.COM (Steve Dyer) (09/06/89)
I'm using a Decstation 3100 as an NFS client mounting several filesystems
on a IBM RT/PC NFS server running AOS 4.3. The 3100 is running Ultrix 3.0
and the RT is running AOS 4.3, Sep '88. Many, but not all, 3100 programs
residing on the RT cannot be exec'ed--they ultimately fail with a swap
error (probably can't page them in) as the RT server spews out the following
(repeated many times):
Aug 16 20:42:01 ursa-major vmunix: xdr_bytes: bad size FAILED
Aug 16 20:42:01 ursa-major vmunix: svckudp_send: xdr_replymsg failed
Now, I could start digging myself, but I wondered if anyone else has
encountered this. Project Athena uses RTs as NFS servers and they behave
flawlessly--their AOS code is not exactly the AOS 4.3 Sep 88 release,
but the IBM NFS code hasn't changed much, I think. So, I'm a bit inclined
to think the problem is with the 3100. Both machines are on the same
thinwire ethernet, and there are no noticable problems with input or output
errors or collisions. Before I get up to my elbows, has anyone else
encountered this?
--
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer
dyer@arktouros.mit.edu, dyer@hstbme.mit.eduehrlich@cs.psu.edu (Daniel Robert Ehrlich) (09/13/89)
In article <4527@ursa-major.SPDCC.COM> dyer@spdcc.COM (Steve Dyer) writes:
I'm using a Decstation 3100 as an NFS client mounting several filesystems
on a IBM RT/PC NFS server running AOS 4.3. The 3100 is running Ultrix 3.0
and the RT is running AOS 4.3, Sep '88. Many, but not all, 3100 programs
residing on the RT cannot be exec'ed--they ultimately fail with a swap
error (probably can't page them in) as the RT server spews out the following
(repeated many times):
Aug 16 20:42:01 ursa-major vmunix: xdr_bytes: bad size FAILED
Aug 16 20:42:01 ursa-major vmunix: svckudp_send: xdr_replymsg failed
Now, I could start digging myself, but I wondered if anyone else has
encountered this. Project Athena uses RTs as NFS servers and they behave
flawlessly--their AOS code is not exactly the AOS 4.3 Sep 88 release,
but the IBM NFS code hasn't changed much, I think. So, I'm a bit inclined
to think the problem is with the 3100. Both machines are on the same
thinwire ethernet, and there are no noticable problems with input or output
errors or collisions. Before I get up to my elbows, has anyone else
encountered this?
--
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer
dyer@arktouros.mit.edu, dyer@hstbme.mit.edu
We have seen this a number of times here also. It has to do with NFS V4 not
being entirely compatible with NFS V3 (which is what is in AOS 4.3). The best
bet is to get IBM to release, productize, or whatever they call it, the latest
version of NFS under AOS 4.3. Of course this might be considered development
and we have been told that AOS is no longer being developed.
--
Dan Ehrlich <ehrlich@shire.cs.psu.edu> | Disclaimer: The opinions expressed are
The Pennsylvania State University | my own, and should not be attributed
Department of Computer Science | to anyone else, living or dead.
University Park, PA 16802 |