Hi Dan, We run our 1Gb ports on the 1GbE switch at 100 Mb (Full Duplex) for just this very reason.
I'd like to try and work out what could be sending the shutdown message to the server. What might be the best way to catch that in some sort of log message or via the Xport? Cheers, Andrew On 16/03/2015, at 12:10 PM, Dan Werthimer wrote: hi andrew, roach1 ethernet ports do not work reliably at 1Gbit/sec. do you have an old 100 Mbit switch you can use? i have no idea why this would cause a board to shut down though. best wishes, dan On Sun, Mar 15, 2015 at 4:42 PM, Andrew Jameson <[email protected]<mailto:[email protected]>> wrote: Hi all, We have a group of 13 ROACH1 boards that have been running stably for 3 years. Recently the 1GbE and 10GbE switches that connect them to some servers was changed over (replaced with identical models and configs). One of the ROACH boards now exhibits some rather intermittent (every few hours) behaviour whereby it initiates a user shutdown whilst running for no discernible reason. The roach_monitor.py script (and roachnest) indicates the shutdown reason is 3 (User Shutdown) And when I happened to be logged into the ROACH via ssh during one of these events, the messages Broadcast message from root@roach (Sun Mar 15 07:38:31 2015): The system is going down for system halt NOW! were broadcast to the terminal. Upon restarting the board, all the temps and voltages are within limits. I'm at a loss about the cause and what might be the next steps in debugging this board. Any advice from the casper world? Thanks, Andrew -- Andrew Jameson Systems & Software Engineer Centre for Astrophysics & Supercomputing Swinburne University of Technology -- Andrew Jameson Systems & Software Engineer Centre for Astrophysics & Supercomputing Swinburne University of Technology +61 3 9214 5612
