On Sun, Mar 31, 2019 at 10:36:53AM +0100, Darac Marjal wrote: > > On 30/03/2019 02:20, Jason wrote: > > I am trying to set up an NFS export on a Ras Pi (Raspbian Jessie). > > Something seems to be missing and I need help troubleshooting it. > > > > At first the NFS port 2049 was not even being opened. After much futzing > > around I discovered that if I manually start rpcbind with > > > > $ sudo /etc/init.d/rpcbind start > > This starts the service manually, but doesn't mark it for automatic > start. To do that, try: > > $ systemctl enable --now rpcbind I get this: systemctl: unrecognized option '--now'
Running it without --now works but it still doesn't survive a reboot. Synchronizing state for rpcbind.service with sysvinit using update-rc.d... Executing /usr/sbin/update-rc.d rpcbind defaults Executing /usr/sbin/update-rc.d rpcbind enable Anything else to try? > > > then restart the nfs-kernel-server I can get it to work, but only if I > > spell out the full ip address of the client. For instance if I put this > > in /etc/exports it works: > > > > /srv/export 192.168.1.1(rw,sync) > > > > but not if I change it to this to allow other machines on the LAN to > > mount (I get: mount.nfs: access denied by server while mounting > > 192.168.1.15:/srv/export): > > > > /srv/export 192.168.1.*(rw,sync)7 > > Are you running "exportfs -rav" or similar after editing that file? > Yes, that exact command. As I wrote in another message, using 192.168.1.0/24 does work, so this is now mainly a curiosity question. > > > > So I have two questions: > > 1) Why does rpcbind not start automatically on boot? > > > > 2) Why can I not specify a range of client addresses in /etc/exports ? > > > > Thanks, Thanks for the suggestions. -- Jason