I've been having problems for some time now with amd under
NetBSD-1.4.1/{i386,sparc}.  The symptom will be that the machine seems
unresponsive to commands, but I can still move windows around just fine.
Running as root, I can do some things.  I believe the problem is related
to an NFS server (solaris) which goes down, then comes back up and somehow
amd doesn't remount it right.  Then the system just sits there waiting to
access the automounted directories.  In fact, if I try 'amq -m' I see:

root# amq -m
amq: failed to contact portmapper on host "localhost". RPC: Timed out

as far as my setup,
/etc/hosts.allow has:

ALL: @my-netgroup, my.host.name, 127.0.0.1, a.trusted.machine

/etc/hosts.deny has:

ALL EXCEPT in.rexecd,sshd:ALL: banners /etc/tcpd/banners
in.rexecd:ALL

in /etc/rc.conf:
portmap=YES             portmap_flags="-l"      # -l logs libwrap

nfs_client=YES                                  # enable client daemons
                        nfsiod_flags="-n 4"
nfs_server=YES                                  # enable server daemons
                        nfsd_flags="-tun 4"


amd=YES                 amd_flags="-l syslog -x error,noinfo,nostats -F
                        /etc/amd/amd.conf"
                        amd_dir=/amd                    # mount dir
                        amd_master=/etc/amd/master      # master map


and my /etc/amd/amd.conf is identical to what we have on some solaris
machines which are running

am-utils version 6.0.1 (build 1).


On the NetBSD machines, I'm running the am-utils that I updated from the
NetBSD source tree about a month ago.

 

I run a very similar setup at home (on all NetBSD-1.4.1 machines), but
have not experienced any of these problems.


Any thoughts?  Is my setup broken?  A broken solaris server?  A bug on my
NetBSD machines?

Thanks for any suggestions.

-Dan