Hi, On Mon, Nov 25, 2013 at 07:32:09PM +0400, Valery Zaporozhchenko wrote: > Hello Eberhard, > > my zx2000 has a year history of such records, but salinfo says that this > is due to machine check which is usually supposed to be a hardware issue. > If people here can read hardware logs you may get them with a null-modem > cable without booting the system. > > Valery Yes, this is likely an MCE (Machine Check Exception)... there should be LEDs going red somewhere (if the machine resembles other HP-UX boxes) and the pattern of LEDs would point to a specific fault condition, usually too generic to be useful in cases like this. This is not only caused by hardware, though - it can also be caused by software doing lowlevel things that the hardware does not appreciate. I've had a couple MCEs that were caused by officially incompatible software that did run, but triggered a nasty in some corner cases. This behaviour - loading elilo, then crashdump and restart - is the same that I get when booting debian 7 in HPVM on my zx6000/HPUX. I have a crashdump, but it's not readable by any tool that I know, unfortunately... Lennert
Attachment:
signature.asc
Description: Digital signature