[Discuss] L1-A in Linux, 10GBE woes

[feenberg at nber.org (Daniel Feenberg)]

On a SUN box one could press L1-A to pause the machine, and then restart 
it later with "go". We have 5 Linux machines connected to a new Brocade 
10GBE switch and wish to update the firmware on the switch. That will take 
about 5 minutes according to tech support, and our Linux boxes (which boot 
from the network) will crash if deprived of / for that long. Some of them 
have long running jobs I was hoping not to interuppt. So I was hoping that 
there might be a way to "pause" them while the switch got updated.

The reason for the update is a (perhaps vain) hope that a problem with the 
switch can be solved. It seems that for some of the computers, after a 
reboot of the computer the switch does not recognize that there is 
connection to that computer until the switch itself is rebooted. Without a 
reboot of the switch the server reports "PXE-E61 Media test failed" while 
PXE is attempting to fetch the boot code and the light on the switch port 
does not go on. If the switch is then rebooted, the light goes on and the 
boot succeeds. That is, it seems that the switch port does not configure
properly if it has not cable attached and connected to a powered on NIC.

Any ideas for either of these problems?

daniel feenberg