Network is configured statically over wired network, kernel is 3.14. rpc.mountd is started with "-d all" as through the conf.d/nfs configuration so that messages should go to syslog-ng's /var/log/messages. nfs is in runlevel default Client ins Fedora 20. NFS is running after the system has booted. However, any attempts to mount the share from the client fail with "Access denied". rpcinfo -p <server> reports nfs v3 and v4 running as it should. There are absolutely no messages indicating any client attempting to connect in the syslog. After /etc/init.d/nfs restart, everything works normal until the next reboot. Reproducible: Always
Please post your `emerge --info' output in a comment.
Created attachment 391364 [details] emerge --info output
After enabling logging RC Start, I have the following in the rc.log * Exporting NFS directories ... exportfs: Failed to resolve p01.fir.local /etc/init.d/nfs: line 48: 1754 Killed ${exportfs} -r [ !! ]
(In reply to Cedric Sodhi from comment #3) exportfs timed out because your network wasn't fully up. the init script then killed it to prevent the system from hanging. you need to make sure the network is fully functional before launching nfs.
Couldn't NFS keep "retrying" in the background? Network is in rc default as it would be on pretty much every normal box, but this surely isn't the normal behaviour one would like to have. So I propose that either something is wrong with the rc dependencies or the script should rather (and preferably) keep spinning in the background and try to resolve until the network is up. In any case, the current situation can hardly be called "working fine" for a pretty normal setup, in my opinion.