Answer:
It was some memory address thing that isn't quite a bug, but that can
slow performance. It appears to be fixed in sander with AMBER9.
I think the addressing issue is more general (is that right?), but the
error message is an Itanium thing. Here's a page for the curious:
http://h21007.www2.hp.com/dspp/tech/tech_TechDocumentDetailPage_IDX/1,1701,161,00.html
I still don't know why it happened when no one appeared to be running
sander. Maybe there was a cron job or intermittent background user
thing going on somewhere. But, it doesn't seem to be doing that
anymore.
:-) Lachele
On 11/27/06, Lachele Foley (Lists) <lf.list.gmail.com> wrote:
> I've been getting many of these unexpected messages on the console.
> They also get written to dmesg:
>
> sander8(30010): unaligned access to 0x60000fffffffa404, ip=0x20000000001df021
>
> No one is running sander on the machine. No one has run sander on it
> recently. I just got 8 more while I was off searching archives, etc.
> If I do a ps on sander, I don't get anything, though I suspect the
> number in parentheses above is a process ID. A routine ps gives:
>
> # ps -ef | grep sander
> root 30803 17262 0 16:13 pts/2 00:00:00 grep sander
>
> The system is RHEL AS 4 U 4 ia64. I do have a bad HD on one side of a
> mirror RAID, but it seems weird that the two would be related.
>
> Any ideas?
> --
> :-) Lachele
> Lachele Foley
> CCRC/UGA
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
--
:-) Lachele
Lachele Foley
CCRC/UGA
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Dec 06 2006 - 06:07:32 PST
