[mythtv-users] strange NFS problem

[Jack]

Hi everyone 

Finally got my Mythbox working.
Very impressive piece of work!

Using Myth 0.14 on Fedora 1
Fileserver runs freebsd 4.8

I store my recordings using NFS.
Mounting line in /etc/fstab :

192.168.2.111:/muziek1  /mnt/nfs                nfs    
udp,rsize=8192,wsize=8192,hard,nfsvers=3

This works fine with normal file copying...

But NOT when using Myth:
When A scheduled recording ends my network stays under heavy load.
(10 base T 80 -90%). I checked and Mythbackend says 'changing from
recordingonly to none. 

BTW Recorded file is fine!

This heavy network activity stops when I close Mythbackend.
Then everything is back to normal.

I used Ethereal to do some sniffing, and this is wat it sees:
192.168.2.104 = MythMachine
192.168.2.111 = BSD server

First during NORMAL recordings.

0.007241 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=1480)
0.007243 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=2960)
0.007244 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=4440)
0.007246 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=5920)
0.007247 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=7400)
0.010877 192.168.2.111 -> 192.168.2.104 UDP Source port: nfs 
Destination port: 800
0.019366 192.168.2.111 -> 192.168.2.104 NFS V3 WRITE Reply (Call In 4)
Len:8192 UNSTABLE
0.024929 192.168.2.111 -> 192.168.2.104 NFS V3 WRITE Reply (Call In 11)
Len:8192 UNSTABLE
0.031791 192.168.2.104 -> 192.168.2.111 NFS V3 WRITE Call, FH:0x0589ad16
Offset:551989248 Len:8192 UNSTABLE[Unreassembled
Packet]
0.031798 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=1480)
0.031799 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=2960)
0.031802 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=4440)
0.031805 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=5920)
0.031807 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=7400)
0.031879 192.168.2.104 -> 192.168.2.111 NFS V3 WRITE Call, FH:0x0589ad16
Offset:551997440 Len:8192 UNSTABLE[Unreassembled
Packet]
0.031881 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=1480)
0.031883 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=2960)
0.031886 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=4440)
0.031888 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=5920)
0.031891 192.168.2.104 -> 192.168.2.111 IP Fragmented IP protocol
(proto=UDP 0x11, off=7400)
0.041478 192.168.2.111 -> 192.168.2.104 NFS V3 WRITE Reply (Call In 20)
Len:8192 UNSTABLE
0.046993 192.168.2.111 -> 192.168.2.104 NFS V3 WRITE Reply (Call In 26)
Len:8192 UNSTABLE

And then AFTER recording has finished:

116.897668 192.168.2.104 -> 192.168.2.111 NFS V3 READ Call,
FH:0x0589ad16 Offset:34922496 Len:8192
116.898309 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=5920)
116.899094 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=7400)
116.900325 192.168.2.111 -> 192.168.2.104 NFS V3 READ Reply (Call In
72487) Len:8192[Unreassembled Packet]
116.901556 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=1480)
116.902789 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=2960)
116.903331 192.168.2.104 -> 192.168.2.111 NFS V3 READ Call,
FH:0x0589ad16 Offset:34930688 Len:8192
116.904020 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=4440)
116.905253 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=5920)
116.906041 192.168.2.111 -> 192.168.2.104 IP Fragmented IP protocol
(proto=UDP 0x11, off=7400)
116.906223 192.168.2.104 -> 192.168.2.111 NFS V3 READ Call,
FH:0x0589ad16 Offset:34938880 Len:8192

This doesn't stop :-(

It gets even stranger:
When I decide to stop an ongoing recording by hand, there is no problem.
Networkactivity stops.

Does anyone know this problem?
It seems Myth related because copying a file works fine?
Maybe there's a sort of a work around, restart the backend if a
recording finished?

TIA Jack