Cause of a system crash

[Zahid Ghani]

We have DS20 with VMS 7.3-2. It crashed with no crash dump. The operator was asked take the CD out of the drive around the time when system crashed.
Is there any way to find out reboot button was pressed.
There is no indication in errlog. Operator log just stops in mid sentence and the system initialisation messages start appearing.

[Hein van den Heuvel]

Hmmm, could this have been an 'ooops' event?
Maybe the operator fat-fingered the cd eject button and hit 'halt' in the process'? Is this a rack-mount? If not, might the cables accidently have been jiggled and a loose connection discovered?
Is the console for the box attached to a monitor of some sort? It may have an indication of what went wrong in its log?
Writes to operator.log require VSM to be alive. If it is down to the >>> console prompt, only that console will have indications.

fwiw,
Hein.

[Zahid Ghani]

Hein,
I belive it was an 'oops' incident but the operator is adamant he did not press the wrong button.
Earlier ,when I said Operator log I meant console log.
I am pretty sure its operator. The only indication is when you have a crash the console messages are different than when you do a powercycle or reset.

As the server is in same cab as another node and every component(power, networks cards etc) is duplicated I don't think it was anything else.

I was wondering if there was any place that might register crash event.

[Volker Halle]

Zahid,

I would not call this a 'system crash', but an (unintentional) system restart.

Even if you have the full console output (you were referring to 'console log') from OPA0:, you may not be able to tell, whether the system restart was due to a 'power fail' or due to someone (or something) triggering the RESTART signal (e.g. by pressing the RESTART button). If the HALT button would have been hit, you would have seen a CPU halted message in the console log.

Volker.