Thanks to all the people who replied. I have found my solution. The
first person with the exact right answer was Jyrki Havia
<havia@cc.helsinki.fi>:
> Have you added those new clients recently? It SEEMS, that
> rpc.mountd needs to be restarted, if new clients are added
> to /etc/hosts on server.
This was a piece of info that I left out of the original question. I
had recently *re-installed* this workstation as a diskless client in
an attempt to re-install the os over the net. In add_client I named
it using the short hostname (I'll never do that again!). Add_client
had changed /etc/hosts to the short name for me (:-) and I then re-made
the NIS maps and exported all the file systems again (naturally).
Later when I was trying to solve the problem I corrected my mistake in
/etc/hosts, re-built the NIS map, etc. etc., but did *not* restart
rpc.mountd after changing /etc/hosts. I had restarted it earlier.
Oh, well.
Most of the respondents hit upon the fact that the server was using
the short name instead of the full name;
Other suggestions
the client must be in /etc/hosts (and correctly, too) not just
known to DNS, and the full name should be the first entry on
the line
(it was after I made the correction)
check that the NFS CLIENT option was turned on in the kernel
(it was)
general NIS problems must be fixed
(NIS seemed to be working fine)
check the /etc/hosts.equiv and /.rhosts files
(that didn't seem to help)
check to make sure there is a hostname.le0 file on the client
(there was)
did the IP address change?
(no)
Many thanks to:
Adam Hammer <hammer@cs.purdue.edu> {who spent time on the phone with me}
Andrew Luebker <aahvdl@eye.psych.umn.edu>
Daniel Quinlan <danq@lemond.colorado.edu>
Danielle Sanine <danielle@systems.caltech.edu>
Dawn Endico <dawn@math.wayne.edu> {who also told me about gopher - thanks}
Jim Davis <jdavis@noao.edu>
Joseph C. Grant <grant@iva.k8.rt.bosch.de>
Jyrki Havia <havia@cc.helsinki.fi>
Mike Higgs <mike@scomp.lancashire-poly.ac.uk>
Mike Raffety <miker@il.us.swissbank.com>
Mike Scott scott@ncifcrf.gov
Roy S. Hallquist, Jr. <rhallqui@rigel.uark.edu>
ray@isor.vuw.ac.nz
ORIGINAL QUESTION:
> I have a Sparcstation II (macaulay) which, due to a corrupted file
> system, just got SunOS 4.1.1 reinstalled on it.
>
> Now, unlike before, I cannot mount filesystems from the server. It is
> installed as a standalone ws. I have rcp'd the standard fstab from
> the server. I can rlogin to macaulay from the server. Everything
> looks in order on the server: /etc/exports has the file systems
> exported to netgroups mathneta and mathnetb, mathneta contains
> macaulay.math.purdue.edu, yp is running and ypmatch mathneta netgroup
> returns the expected list, rpc.mountd is running (and I have killed
> and restarted it just in case), and nfsd (8 of them) are running.
> Other standalone clients have these same filesystems mounted ok.
>
> Problem:
>
> I get "access denied . . . giving up on..." messages for everything I
> try to mount from the server on macaulay UNLESS I take the access=...
> option off of the exports line. I.e. when exportfs is exporting to
> the world macaulay can mount the filesystem, otherwise it can't.
>
> Can anyone suggest what I might be missing here? I need to get this
> workstation going soon.
>
> Thanks a million,
> Steve.
>
> ====================================================================
> Steve Holmes \ Bitnet: sjh%math.purdue.edu@purccvm
> Systems Administrator \ Internet: sjh@math.purdue.edu
> Dept. of Mathematics \ uucp: ...!purdue!sjh
> Purdue University, \ phone: (317) 494-6055
> W. Lafayette, Indiana 47907 \ Fax: (317) 494-0548
> ====================================================================
Steve.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:53 CDT