NFS responding on wrong interface

[tom marinis]

Greets Bill,

--- Bill Campbell <linux-sxs at celestial.com> wrote:
> I'm having a wierd problem at one of our customer's sites. 
> The system has
> multiple IP addresses bound to one NIC, eth0 and eth0:[0-3]. 
> NFS clients
> can't mount NFS directories because the UDP replies are coming
> from
> addresses other than eth0.
> 
> I've looked at the source for mountd.c, and there's no option
> to bind to a
> specific interface (the man pages don't have one so I went to
> the source to
> make sure that there's not an undocumented option).  The
> system in question
> is running Caldera eDesktop 2.4, but I've looked at the source
> on SuSE 8.2
> Professional and the code is largely the same.
> 
> It seems to me that the return packet should show as coming
> from the
> primary interface on that network.  The only thing I see that
> looks a
> little strange is that ``netstat -rn'' shows two routes to the
> internal
> network, both on eth0:
> 
> # netstat -rn
> Kernel IP routing table
> Destination     Gateway         Genmask         Flags   MSS
> Window  irtt Iface
> 192.168.254.0   0.0.0.0         255.255.255.0   U         0 0 
>         0 eth0
> 192.168.254.0   0.0.0.0         255.255.255.0   U         0 0 
>         0 eth0
> 127.0.0.0       0.0.0.0         255.0.0.0       U         0 0 
>         0 lo
> 0.0.0.0         192.168.254.8   0.0.0.0         UG        0 0 
>         0 eth0
> 
> Any ideas?
> 
> Bill


I know I'm really late with this reply, but if
you are still trouble shooting, I do have a few ideas, 
but you'll probably laugh at me for mentioning them, since
I am still a newbie.


I noticed you've got your multiple ip's in the 192.168.254.X
range for eth0.  Can I ask if maybe the other clients 
moved very recently to differnet network subnets?


The only other things I can think of that you probably already
checked ;


the /etc/export file had been modified with wrong ip
   address or computer names to the proper shared directories,
   and of course, white space;

the mount volumes had been unmounted for fsck maintainence,
  with the clients still on the mount;

maybe someone modified the /etc/hosts.allow file, on either the
  server or the client, without remembering to tell anyone,
  for reasons of security.  A DENY ALL statement maybe
  somewhere?



I'm sorry Bill, but that's all I can really think of.   That's 
all this newbie can think of...
 


__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com