RPCPROG_MNT: RPC: Timed out / receiving NFS error when trying to mount NFS file system after make world

[Rick Macklem]

On Tue, 2 Jun 2009, Rick Macklem wrote:

>> 
>> 
>> I'm having troubles with nfsroot on avr32 after updating my dev box to HEAD 
>> on May 31.
>> 
>> I can see MNT RPC packet coming from the avr32 board and running mountd -d 
>> shows "mountd: mount successful" when the packet is received but no answer 
>> is transmitted, rpc is sent with udp.
>> I can mount the same export from my osx workstation when using tcp.
>> 
> Ok, I've poked at it a little more and the case that seems to be broken
> is the "-h nfs-server.cis.uoguelph.ca" option on nfsd. Without "-h" or
> with "-h 131.104.49.243" it seems to work. (This affects udp but not tcp.)
>
> I haven't yet figured out why that case is broken, but I'll keep fiddling
> with it. (For me the getaddrinfo() fails for this case.)
>
> If you are not using the "-h" option on nfsd and udp isn't working, I
> haven't got an explanation, because it seems to work for me?
>
I typed the above (and the bit about /etc/exports continuation lines)
before I had poked around with it enough. The continuation lines seem
to work and the "-h nfs-server.cis.uoguelph.ca" message logged is just
because I'm using ipv4 only.

I seem to get udp mounts to work fine, but...

I've seen what Robert mentioned. I had just assumed it was some weirdness
in my local lan. I've found that, if you "ping <server>" before doing the
mount, it seems to always work. (I usually see it when mounting a 
Solaris10 client to the FreeBSD8 server and it eventually times out and
retries successfully. I normally use tcp mounts, so I didn't "connect"
that with this problem.)

So, maybe it's some network interaction issue and not an obvious goof up
by me w.r.t. changes in the utilities? (The changes I did shouldn't have
had anything to do with the mount protocol code in them.)

rick