Hi Folks, My net6501 had been humming along happily since it arrived in 06/2012 .. I don’t often need to restart it - the last time was 91 days ago following a protracted power failure my UPS ran out of juice.
At any rate, last night I noticed that a file system had overflowed and so deleted a bunch of files and restarted a few processes, the file space didn’t come back straightaway and so I thought about restarting a few more things but then it was late and had been a long day and so I thought I’d simply reboot it. That was sort of a bad decision, as 4 hours later I finally got to bed. Here’s what happened. I rebooted and the board’s led’s cycled as normal but then only the power and error led remained lit. The console was catatonic. My first thought was that the attached SSD had calved, and so powered off, unplugged that, powered on. Same result. I then brought the unit over to my lab bench and hooked it up to my variable power supply. Set the power balance at about 12v and powered on. Same. Held the reset button down until the unit powered off. Released and held it in again until the unit powered back on. Same result. At 12v I then tried all of the console baud permutations - 9600, 19200, 38400, 57600 and 115200 to make sure that I wasn’t doing anything silly. Still no output, and just the two leds - the power and the red error. Grabbed a regular 2.5” laptop drive, plugged it in to the net6501 and powered on again. The drive spun up, no problem. Took the SSD used in the failed net6501 and plunked it on my ‘backup’ net5501 via an external eSATA enclosure and up the net5501 came - config was a bit upset about the change in interface - em0 -> vr0 and disks da0->ad0 but otherwise perfectly functional. Phew. With internet access restored courtesy of the net5501 I looked for net6501 error and found a thread on this list that recommended ‘+++’ on the serial port to get to control mode or somesuch. No output at the any of bauds, with complete power off in between each test. The cpu was warm but not hot, normal really, and the board was drawing about 0.3 amps at 15v, but it was clear that the network ports were not powering on. First I tried a ge loopback plug I’d made a while back in each port - none fired up. Then I tried cabling each port in turn to a gigE switch and neither side (the switch nor the board ports) lit up. So now what? Well. I started ramping down the voltage, and the current increased as expected. But then something magical happened at 5v, 3 of the network leds (the three that had been in use prior to the failure) lit up. Still nothing from the link leds though and still no console at any of the bauds, nay any effect of ‘+++’. And so that's where I’m at. Kind of out of ideas, and not sure of how much more I can diagnose with this little fellow. I suspect that something in the BIOS power-on sequence has failed. The strange thing (and what gives me a bit of hope) is that the board was running perfectly until had decided on the reboot. Perhaps the BIOS has become corrupted? Does anyone know if that’s possible and if so if there’s any way to fix it - maybe a cheap and cheerful JTAG jig? Any and all suggestions welcome :) Cheers, —Andrew _______________________________________________ Soekris-tech mailing list [email protected] http://lists.soekris.com/mailman/listinfo/soekris-tech
