Hi Matt, We've got a roach1 configured as a spectrometer and have similar issues.
Note that the FPGA part of the roach and also the Katcp communications seem to work well. The symptom I have is ssh is not reliable without a reboot. We've left the roach running as a spectrometer for a month or more without cycling power and it seems to provide a steady stream of good data at fairly high rates, with little trouble. Only the ssh root@roach just hangs, while ping does respond. As I said, after cycling the power, ssh is OK. Glen > Hi, > > Is anyone willing to explain what seems to me to be an unexpected > behavior in a windows+cygwin PC to Roach1 network connection ? > > With the Roach1 in the uboot state. Network address (192.168.3.240) > acquired via dhcp, from a directly connected windows+cygwin > PC host system (192.168.3.2). > > Roach1 can ping the PC exactly as expected. > PC cygwin ping of Roach1 hangs. > > While PC ping is hung if the Roach1 pings an unused address > say 192.168.3.3 then the PC pings of the Roach1 un-hang and > return just as expected. When the Roach1 pings give up > after the N failures the PC pings return to the hung state. > > I've only tried this on 1 particular Roach1 but it is > completely repeatable across power cycling, cable reseating > and resetting the Roach1 and anything else I could think > to try. > > When booted to linux this Roach1 responds to pings exactly > as expected. > > Maybe to be expected because Roach1 uboot doesn't respond > to some particular type of packet requests the PC requires ? > > Thanks, > Matt >
