I have created two nfs mount points on my debian box and I can
see them (as root) on my gentoo box with "showmount -e 192.168.0.3"
(that's the debian box'es net addr).

If I try and mount one of these exports the mount command hangs
and I have to kill both it and the bash shell to recover the
xterm or the vt login. A ps ax doesn't seem to show portmap
running, that may be my problem. If so How do I start portmap?
I DO seem to have portmap installed.

ken enlightened us with:
> A ps ax doesn't seem to show portmap running, that may be my
> problem. If so How do I start portmap?

Make sure you have done

$ rc-update add nfs default

Then do /etc/init.d/nfs start.

Sybren
--
The problem with the world is stupidity. Not saying there should be a
capital punishment for stupidity, but why don't we just take the
safety labels off of everything and let the problem solve itself?

Sybren Stuvel wrote:

> ken enlightened us with:
>> A ps ax doesn't seem to show portmap running, that may be my
>> problem. If so How do I start portmap?
>
> Make sure you have done
>
> $ rc-update add nfs default
>
> Then do /etc/init.d/nfs start.
>
> Sybren
Thanks, that seems to have started portmap. BTW can you (or someone) point
me to a url with good documentation on managing Gentoo's rc files? This is
the kind of stuff I am always scratching my head about when installing new
system pacakges.

ken enlightened us with:
> BTW can you (or someone) point me to a url with good documentation
> on managing Gentoo's rc files? This is the kind of stuff I am
> always scratching my head about when installing new system pacakges.

Just do 'rc-update' without any arguments. Also check out rc-status.
Google for them to find more.

Sybren
--
The problem with the world is stupidity. Not saying there should be a
capital punishment for stupidity, but why don't we just take the
safety labels off of everything and let the problem solve itself?

ken wrote:

> If I try and mount one of these exports the mount command hangs
> and I have to kill both it and the bash shell to recover the

Did you use the 'tcp' option when you mounted it? The default is udp, which
didn't work for me.

Blake

On 2005-02-08, Blake <bleverett@pants.att.net> wrote:
> ken wrote:
>
>> If I try and mount one of these exports the mount command hangs
>> and I have to kill both it and the bash shell to recover the
>
> Did you use the 'tcp' option when you mounted it?
> The default is udp, which
> didn't work for me.
>

If nfs/tcp works but normal nfs, it is a typical indication
of a packetloss/NIC problem.
Try with ping (or other better tools ???) if you have
packetloss.


/klacke

--
Claes Wikstrom -- Caps lock is nowhere and
http://www.hyber.org -- everything is under control

Claes Wikstrom wrote:

> On 2005-02-08, Blake <bleverett@pants.att.net> wrote:
>> Did you use the 'tcp' option when you mounted it?
>> The default is udp, which
>> didn't work for me.
>>
>
> If nfs/tcp works but normal nfs, it is a typical indication
> of a packetloss/NIC problem.
> Try with ping (or other better tools ???) if you have
> packetloss.
>
>
> /klacke
>

I don't think it was packet loss. I checked with ifconfig and another
program, and found no bad packets. So I don't think that it was a hardware
problem. Ping -f loses zero packets between the two machines.

The symptom I had was that NFS traffic would slow to a crawl when doing a
big transfer, and only in one direction. I could xfer from my old Mandrake
system to my new Gentoo system just fine (NFS drives exported from the
Gentoo system). But when moving the other way, the file transfer would
work fine for a while (10 secs to 1 minute), then it would slow to a crawl,
going from about 10MB/sec to 40KB/sec. It would be really slow until the
xfer was done. A new xfer would always work at high speed until it went
nutty, then crawl for the remainder of the transfer. The file transfer
always worked, but very slowly.

After a ton of research I saw a suggestion to use the TCP mount flag, which
has made NFS pretty solid for me.

Blake