OpenVMS problems

[zoli]

Mike, please help!

Could you please tell me on English what does it mean:

Processor Detected - BCache single bit ECC error
(followed by segment and bits in hex)
and
Processor correctable error through VECTOR (some hex number)

...this is what I got when I boot alpha.

Is there a memory problem or the CPU is dying.

Thank you.

[miker_alpha]

I _think_ its probably memory, but as soon as I'm at HP I'll ask a HW person. Unfortunately that might only be on Sunday afternoon...

Best of luck!
MikeR

[miker_alpha]

what i could find so far:
it might (small chance) be memory, which you can check by swapping out 1/2 memory at a time
or it might (more likely) be onboard CPU cache memory, (have there been heat problems?) in which case if you haven't got maintanance its probably lost (IIRC needs a MB swap.)

Sorry
I'll try to come back with more information Sunday.

MikeR

[zoli]

Thanks Mike. I'll try with memory swap... and we'll see later...

Good that I was able to build Vim7 two days ago...

[zoli]

I have no idea what was the problem - but once more helped the unscientific approach.

I opened the box, took a vacum cleaner and removed the few years old dust. Touched, moved a bit the memory chips - and rebooted the system.

It worked.

To be honest, after those hex dumps that the cpu produced last night I haven't even imagine that it will be that easy.

But this might be the last round for some parts in alpha box - therefore we have to be prepared for longer and more serious outage - I do not have any spare parts for this box (except disks).

[miker_alpha]

Then it probably was memory:
Sometimes the memory card contacts go "dry" - dust or oxidation; just moving them a tiny bit will reseat the cards.

Whatever it was, good to see Alpha back on the air - It could take another 5 years till it happens again!

MikeR

[miker_alpha]

Showed your errors to our HW person.
According to him its almost certainly the CPU or adjacent on-board BCache. When I told him you re-seated memory, MB etc. and now it works, he said something like "be grateful, you're lucky!"

(he also said there's a small chance it is/was a memory glitch- a one-bit correctable error)
perhaps some of the dust you cleared was conductive?

Sorry I can't give you better news.
Regards,
MikeR.

[zoli]

Thank you Mike... this is good enough.

To be honest, I deeply appreciate your fast and competent reply - not many of polarhome like organizations enjoy direct HP support's hospitality.

Now, I know that we need to look after some alpha box - and need to be prepared for the worst.

[miker_alpha]

There seems to be another/different problem on Alpha now (approx 08:00 GMT, Tues 16th)

The Alpha seems to be up -- i get a response on port 3017 (QOTD), but no response on port 763. Is Apache down?

MikeR

[zoli]

I have no idea what have happened... this must be something with the ISP.