[Summary} Panic/Crash on multi-processor Alpha

From: Jay Grover (Veda) <"Jay>
Date: Fri, 15 Aug 1997 07:31:00 -0400

First off, thanks to everyone who responded, you guys are great!
This is a known problem in OSF/1 v3.2c. I have downloaded the patches
and will apply them shortly. Here are my "best" replys:

Flack, Peter wrote:
>
> Jay,
>
> The CPUs should be installed in the lowest numbered slots of the TLSB
> (slot 0 for first dual CPU board). Memory should be next to CPU(s)
> going up in slot numbers. I/O boards should be in highest numbered
> slots (starting with slot 8) - grant cards should fill in the middle. I
> am not sure what the slots 10 and 11 are that you are referring to - the
> TLSB on the 8400 is a 9 slot bus (0 - 8).
>
> Thanks,
>
> Peter Flack
> System Engineer
> Best Western International
> (602) 780-6759
> flackp_at_bestwestern.com

Martin Moore wrote:
> Installing the latest 3.2c patch kit should prevent this from happening
> again. The easiest way to get this is via our patch web server
> (http://www.service.digital.com:8031). In the selection box, click
> "All databases" to deselect it and click "Digital UNIX" to select it.
> Then search on the word 'aggregate'. You'll see the available aggregate
> patch kits for each version. Click on the 3.2C one and you'll be in
> business.
>
> Martin
> --
> Martin J. Moore 5555 Windward Parkway West
> Digital UNIX Support Alpharetta GA 30004-7407
> Digital Equipment Corporation 1-800-354-9000 x31679
> mailto: martin_at_alf.dec.com DECATL::MARTIN

> > Greetings to all my fellow Alpha managers!
> > A very strange problem is occuring on one of my Alpha servers. The
> > machine is a DEC 8400 Model 5/300 with dual CPUs (type 7000). It will
> > run fine for days at a stretch, but will eventually panic, crash and
> > then reboot.
> > The exact message and sequence of the problem occurs as follows:
> > 1) Message: Panic (cpu10): fill_tlsb_can't_translate address of CPU
> > 2) Almost exactly five minutes later, the machine will crash, then
> > reboot (error event message is recorded before crash).
> > Some additional details (don't know if they are relevant or not!):
> > - The first CPU (master) is located at slot 10, the second CPU is
> > located at slot 11.
> > - Both CPUs are part of the default processor set (0).
> >
> > Does anyone with multi-processor machines have any insight as what
> > this
> > problem is or why it is occuring?
> > Should the first CPU be installed in slot 10? (The machine was
> > configured before I got this job, so I don't have any idea as to its
> > "correctness").
> > I've looked everywhere for a reference to "fill_tlsb" and can't find a
> > thing. Any help would be greatly appreciated. I will summarize.

-- 
Jay Grover
"No matter where you go, there you are."
jgrover_at_mbvlab.wpafb.af.mil
Received on Fri Aug 15 1997 - 13:46:56 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:36 NZDT