[UCLA-LUG] Re: Linux digest, Vol 1 #372 - 17 msgs

Kai-Min Sung nimiak@ucla.edu
Fri, 07 Apr 2000 08:36:37 -0700


One of the jacked up features of knfsd is that the nfsd daemons will
often die without releasing the port it listens on (udp 2049). I
bet if you run 'netstat -a |grep 2049', you'll see

udp        0      0 *:2049                  *:*

in use. That's why you're seeing the "nfssvc: Address already in use"
message below.  What I normally do is reboot the machine to
flush network ports. But, a very unorthodox method of working around
this would be to start nfsd with the -p option (man nfsd) to specify
a port other than 2049 and then on your client end use the mount
option feature port=xxx, where xxx is the port number you have nfsd
running on the server (man mount).

-Kai
nimiak@ucla.edu


>
>
> I do not have /etc/rc.d/init.d/nfslock on this machine.  Also this is
> what I get:
>
> [root@cypress mydoc]# /etc/rc.d/init.d/nfs start
> Starting NFS services:                                     [  OK  ]
> Starting NFS statd:
> Starting NFS quotas:
> Starting NFS mountd:
> mountd: couldn't stat /var/lib/nfs/xtab
>                                                            [FAILED]
> Starting NFS daemon:
> nfssvc: Address already in use
>                                                            [FAILED]
> [root@cypress mydoc]#
>
> That /proc/filesystems I just sent was for both machines.  Thanks!
>