On 11/9/2018 12:59 PM, Jon Elson via cctalk wrote:
On 11/08/2018 11:09 PM, Douglas Taylor via cctalk wrote:
I hadn't started this DEC Alpha 3000-300 since last summer, and
booted it up so I could load the new PAK's the other day.
The result was that it completes almost the entire OpenVMS startup,
but then crashes with the following:
%SET-I-INTSET, login interactive limit = 64, current interactive
value = 0
**** OpenVMS Alpha Operating System V8.4 - BUGCHECK ****
** Bugcheck code = 000001CC: INVEXCEPTN, Exception while above ASTDEL
** Crash CPU: 00000000 Primary CPU: 00000000 Node Name: A300
** Highest CPU number: 00000000
** Active CPUs: 00000000.00000001
** Current Process: DECW$STARTUP
** Current PSB ID: 00000001
** Image Name: A300$DKA0:[SYS0.SYSCOMMON.][SYSEXE]DECW$CONFIG.EXE;1
**** Starting selective memory dump at 8-NOV-2018 21:57...
The disk is a SCSI2SD board with an 8 GB SD card. It had been
running just fine, until now.
Does this crash point to a hardware or software problem?
ASTDEL indicates that something in the OS caused a machine check
interrupt. Does DECW$CONFIG mean it is the DecWindows config process
that bombed? Can you get the OS to boot in non-windows mode, and try
to log in from a serial terminal? If DecWindows is blowing up, it
could be that the graphics adapter has a problem, is set to the wrong
address, or is a different type than the software is expecting. You
might unplug and reseat all mezzanine boards and cycle all DIP
switches to make sure it is not just a bad connection.
Jon
I was able to boot the system using a serial console only by setting the
jumper listed as 'W2' so that pins 2-3 were connected rather than 1-2.
It did complete the system startup and I was able to log in to the
system account.
This 3000-300 has no mezzanine boards, the system consists of a
motherboard with a cpu board the rides on a mezzanine type of
connector. The video comes off the motherboard. I did try wiggling the
memory dimm cards. I'll look and reseat any jumpers that are there.
The last time I tried to bring up the system using only the serial
console I got the bugcheck crash, which had only occurred when I tried
to boot into graphics mode.
It maybe that the DECW$Config.exe image is bad, since that is where the
bugcheck always occurs. Although this is confusing because it did
successfully boot in console mode previously.
So, the choices now seem to be; (1) replace the hardware (I think I have
another 3000-300) or (2) reinstall the software. Don't like either one.
