Re: nfs udp problem

[Alexander Samad <alex@samad.com.au>]

Just to add to this, seems like v4 doesn't work either, when I use
nfsv4 and mount on a client I get userid and gids like this 4294967294
4294967294

bugger

On Sun, Jun 13, 2010 at 3:01 PM, Alexander Samad <alex@samad.com.au> wrote:
> Hi
>
> I am having problems mounting mu nfs shares via udp, if i flip to tcp
> all is well.
>
>
> 14:49:35.415978 IP 192.168.10.221.330032314 > 192.168.11.10.2049: 40 null
> 14:49:35.415978 IP 192.168.10.221.330032314 > 192.168.11.10.2049: 40 null
> 14:49:35.416097 IP 192.168.11.10.2049 > 192.168.10.221.330032314:
> reply ok 24 null
> 14:49:35.416114 IP 192.168.11.10.2049 > 192.168.10.221.330032314:
> reply ok 24 null
>
>
> this is a capture of the packets flow on the nfs machine
> (192.168.11.10) the client has made request
> mount -o vers=4,udp  nfs:/exports/video /exports/video/
>
> the second packet - the reply is not making it back to the laptop - I
> have check with another box on the wire to see if it see the second
> packet but it doesn't
>
> change the udp to tcp and it works.
>
> the reason i have 2 copies of each packet is I am using a bridge, and
> you see the packet from the br0 and from eth0
>
>
> auto br0
> allow-hotplug br0
> iface br0 inet static
>        bridge_ports eth0 eth1 eth2 eth3 vde0
>        bridge_stp on
>        address 192.168.11.10
>        network 192.168.11.0
>        netmask 255.255.255.0
>        gateway 192.168.11.1
>        broadcast 192.168.11.255
>        pre-up ifconfig eth0 up
>        pre-up ifconfig eth1 up
>        pre-up ifconfig eth2 up
>        pre-up ifconfig eth3 up
>        pre-up ifconfig vde0 up
>        down ifconfig eth0 down
>        down ifconfig eth1 down
>        down ifconfig eth2 down
>        down ifconfig eth3 down
>        post-up /sbin/ip -6 addr add dev br0 fe80:11::10/64 || true
>        post-up sysctl -w net.ipv6.conf.br0.autoconf=0 || true
>        post-up /sbin/ip -6 r a default via fe80:11::1 dev  br0 advmss
> 1440 metric 50 || true
>        post-up /sbin/ip -6 a flush eth0 || true
>        post-up /sbin/ip -6 a flush eth1 || true
>        post-up /sbin/ip -6 a flush eth2 || true
>        post-up /sbin/ip -6 a flush eth3 || true
>
> if I change the mount options to vers=3, then I getting a locking time out.
>
> I am stumped - I have tried bringing down the bridge interface and
> just use eth0, but same problem.
>
> I am stuck, I want to use udp, becuase when my machine goes to sleep,
> the nfs just works afterwards, no need to umount and then remount and
>
> Alex
>