[Info-vax] Alphaserver ES47: Suspected broken CPU, unable to stop/cpu 2

[Roy Omond]

On 08/08/18 12:24, abrsvc wrote:
> On Wednesday, August 8, 2018 at 3:09:57 AM UTC-4, Robin Schrievers wrote:
>> On Wednesday, 8 August 2018 08:53:08 UTC+2, Hans Vlems  wrote:
>>> Is it possible to physically remove cpu 3, pull the board?
>>> Hans
>>
>> That will be a challange. As the box is not near me i can't do that myself. We might be able to have remote hands do that, but that would require some very detailed explanation.
> 
> Without getting into details, where is this machine located?  Perhaps there are some of us that are local to the machine that can assist.

I'd guess Wageningen, the Netherlands.

I'm rather surprised nobody's even suggested to actually analyse the
dump file (+ probably the errorlog).  You need to figure out exactly
what has caused the machine check in kernel mode.  It *might* be
purely coincidence that it seems to only ever occur on CPU 2.

Robin, please start by posting the output from:

	$ analyse/crash sys$system:
	SDA> clue crash

Next you'll probably need to:

	$ analyse/crash sys$system:
	SDA> clue err

This will dump the errorlog buffers active at the time of the crash
to a file sys$login:clue$errlog.sys

It will then depend on whether you have DECevent or WSEA installed
on your system in order to analyse this file.  Try first:

	$ dia sys$login:clue$errlog.sys