Small NAS keeps hanging: 9.0 RC3

[James Mansion]

I'm trying to run Windows backup to my NAS, which is running 9.0rc3.

nas# uname -a
FreeBSD nas 9.0-RC3 FreeBSD 9.0-RC3 #0: Sun Dec  4 08:56:36 UTC 2011     
root at farrell.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC  amd64
nas#


I'm using Samba on a ZFS pool.

System stops responding to the network, nothing untoward visible on top, 
pressing the reset button initiates a halt and reboot and it all comes 
back up.

Failure is after 38.46GB of file backup and 138.56GB of system image.  
Backup does not complete - the drive goes away. Putty sessions are 
dropped and sshd is unresponsive.

I get this in syslog:
Dec 17 13:34:49 nas kernel: msk0: watchdog timeout
Dec 17 13:34:49 nas kernel: msk0: prefetch unit stuck?
Dec 17 13:34:49 nas kernel: msk0: initialization failed: no memory for 
Rx buffers
Dec 17 13:38:09 nas ntpd[1685]: sendto(193.110.251.9) (fd=22): Network 
is down
Dec 17 13:39:09 nas ntpd[1685]: sendto(194.71.144.71) (fd=22): Network 
is down
Dec 17 13:41:20 nas ntpd[1685]: sendto(194.109.64.200) (fd=22): Network 
is down

Symptom is always the same.

nas# ifconfig
msk0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
         
options=c011b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,TSO4,VLAN_HWTSO,LINKSTATE>
         ether 00:50:43:00:12:40
         inet 192.168.0.82 netmask 0xffffff00 broadcast 192.168.0.255
         nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
         media: Ethernet autoselect (1000baseT 
<full-duplex,flowcontrol,rxpause,txpause>)
         status: active

dmesg has:

msk0: <Marvell Technology Group Ltd. Yukon EC Id 0xb6 Rev 0x02> on mskc0
msk0: Ethernet address: 00:50:43:00:12:40
miibus0: <MII bus> on msk0
e1000phy0: <Marvell 88E1111 Gigabit PHY> PHY 0 on miibus0
e1000phy0:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 
1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto, 
auto-flow

This is an addin card.

I also have an adapter on the motherboard in the system which I *could* use:

msk1: <Marvell Technology Group Ltd. Yukon Ultra 2 Id 0xba Rev 0x00> on 
mskc1
msk1: Ethernet address: dc:9c:52:07:a5:7f
miibus1: <MII bus> on msk1
e1000phy1: <Marvell 88E1149 Gigabit PHY> PHY 0 on miibus1
e1000phy1:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 
1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, auto, 
auto-flow

And I could swap the card for an Intel CT desktop adapter if its likely 
to be an improvement.  Is it worth trying the other Marvell chip at all?

(This happened with RC2 but I didn't capture the cause.  Uptime with RC3 
had seemed better, but i think that's because I hadn't initiated a full 
backup with system image.)

James