em0 timeout disconnects server

[Jeremy Chadwick]

On Tue, Aug 02, 2011 at 12:27:57AM +0200, Willem Jan Withagen wrote:
> A server just all of a sudden dropped from the network.
> uptime was 26days.
> 
> This got my ZFS server hanging:
> 
> Aug  1 23:39:58 zfs kernel: em0: Watchdog timeout -- resetting
> Aug  1 23:39:58 zfs kernel: em0: Queue(0) tdh = 942, hw tdt = 977
> Aug  1 23:39:58 zfs kernel: em0: TX(0) desc avail = 985,Next TX to Clean
> = 938
> Aug  1 23:43:24 zfs kernel: em0: Watchdog timeout -- resetting
> Aug  1 23:43:24 zfs kernel: em0: Queue(0) tdh = 147, hw tdt = 163
> Aug  1 23:43:24 zfs kernel: em0: TX(0) desc avail = 1006,Next TX to
> Clean = 145
> 
> ifconfig down/up did not fix anything, un/plugging the ethernet did not
> do anything either. rebooting did fix it.
> 
> Serious maintenance jobs only starts after 0:00.
> 
> --WjW
> 
> uname -a:
> FreeBSD zfs.digiware.nl 8.2-STABLE FreeBSD 8.2-STABLE #10: Wed Jul  6
> 21:57:36 CEST 2011
> root at zfs.digiware.nl:/home/obj/usr/src/src8/src/sys/ZFS  amd64

Please provide "dmesg" output pertaining to the NIC (dmesg | grep em0
would be sufficient).

> pciconf -lv:

Please re-run this with -lvcb and include only the Intel NIC (em0).

Also please provide output from command "sysctl dev.em.0".

Thanks.

-- 
| Jeremy Chadwick                                jdc at parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                   Mountain View, CA, US |
| Making life hard for others since 1977.               PGP 4BD6C0CB |