DaZZa wrote:

The Cisco device would block the bad port if it detects a problem.

The switch ports in the c8?? are too dumb to do that.


It's probably a jabbering network interface card (eg, sending
the last packet repeatedly, with corruption). These are usually
isolated by the switch, but we're talking household kit here
(which tends to do cut-through switching which will pass jabbering
frames whereas enterprise switches usual take the slower but
safer path of receiving and checking the entire packet before
re-transmitting it).

So the c8xx is getting an interrupt per incoming faulty frame
and is locking up because it's spending all of its time in
the Cisco equivalent of kernel system call mode.

Next time this happens have a look at the Transmit light
on the NIC or the Receive light on the switch port the NIC
is in.  If it's solid report a bug to netdev (with ethtool
output) and toss the NIC.

--
 Glen Turner         Tel: (08) 8303 3936 or +61 8 8303 3936
 Australia's Academic & Research Network  www.aarnet.edu.au

--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Reply via email to