[plug] NFS problem

Matt Kemner zombie at networx.net.au
Tue Dec 1 08:57:05 WST 1998


On Mon, 30 Nov 1998, Nick Bannon wrote:

> 1. Sometimes actually stopping and starting the daemons in question (eg
> Debian "/etc/init.d/netstd_nfs stop", "/etc/init.d/netstd_nfs start")
> rather than just SIGHUP-ing them ("/etc/init.d/netstd_nfs reload")
> helps.

Occasionally you need to restart the portmapper too, which you do by
killing the nfs daemons as above, running
/usr/sbin/pmap_dump > /root/table.tmp
killall portmap (or rpc.portmap or rpc.portmapper or whatever your system
calls it)
/sbin/portmap or /usr/sbin/portmapper, or whatever.
/usr/sbin/pmap_set < /root/table.tmp
rm -f /root/table.tmp

> 2. "/usr/sbin/showmount -e nfsserver.domainname.foo" is very handy for
> telling you whether you've a client problem or a server problem. If
> it tells you "RPC: Program not registered", then one of your daemons
> isn't running.

9 out of 10 times when I get this error I've not configured the
hosts_access(5) files for portmap (which needs it's own special entry, as
per portmapper.txt that comes with the distribution -
/usr/doc/netbase/portmapper.txt.gz on debian)

 - Matt



More information about the plug mailing list