Portmap startup: stumped

Hello

I have tried several hours to resolve the following seemingly simple
problem. I recently effected some security and optimisation changes on
a linux box. I thought I was being diligent, bouncing services and
testing as I went along. Then I did a reboot...

While booting up, everything looks normal until NFS mountd is invoked,
whereupon my poor box hangs for some time before reporting

Starting NFS mountd: Cannot register service: RPC: Unable to
receive; errno = Connection refused

No problem, I think: portmapper isn't running for some reason (I
changed something that affected portmapper? hmm....). Indeed, rpcinfo -p
throws an error, and ps -aux confirms. I then double checkconfig
(ha ha) portmapper but that's fine. Furthermore starting
portmapper as root and bouncing NFS yields expected results

# /sbin/portmapper
# /etc/init.d/nfs restart
# rpcinfo -p

NFS and ancillary services all present.

So this all boils down to the fact that portmapper won't start
properly on its own. It's as though it isn't starting as root.

/var/log yields a "portmap: cannot bind upd: address already in use". Something
tells me I'm missing something...

I have read a slew of other postings with very similar symptoms to no
avail. Note that from the perspective of network clients, NFS services
seem unaffected (i.e. I can manipulate exported shares). So the only
real problem is the long bootup time, and my now frustrated curiosity
to know what's going on here.

Best regards

Edward Garson

Edward G wrote:
>
> /var/log yields a "portmap: cannot bind upd: address already in use". Something
> tells me I'm missing something...
>

This tells that something is already using the portmapper
port (111/UDP). Use (as root) 'netstat -upan' to check.

HTH

--

Tauno Voipio
tauno voipio (at) iki fi