[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: "Hard memory error" -- where?
-----BEGIN PGP SIGNED MESSAGE-----
On Aug 13, 2008, at 5:09 AM, Hauke Fath wrote:
my trustworthy ss20 crashed yesterday with
Aug 12 19:19:12 pizza /netbsd: store buffer copy-back failure at
Aug 12 19:19:12 pizza /netbsd: memory fault: type
Aug 12 19:19:12 pizza /netbsd: panic: hard memory error
Aug 12 19:19:12 pizza /netbsd: syncing disks...
-- I had extended it with a few more memory modules a month or
three ago. It wasn't even hot that day, and the machine does not
have any internal drives; 448 MB of RAM though, and 2x SM71.
Is there any way to learn from the panic message _which_ module
displays the hard error?
Not without the fault's physical address.
Or any memory test tool that would provide module information?
Yes, the firmware will do just that.
IIRC all you need to do is to run 'selftest' on the /memory node. I'm
sure there's an alias for it ( test-memory I think ) but I'm not 100%
sure of its exact spelling ( might be memory-test or memtest or
something - try the help command ). Also, I'm not sure if you need to
change the memtest-megs# PROM variable ( I think it only tells the
firmware how much memory to test on startup but I'm not sure if it
has any influence on memory checks run by hand ).
Most OBP versions will tell you outright which slot contains the
faulty module. Some others will at least give you a physical address,
you may need the SS20 Service Manual to translate that into a slot
number ( was available from sun.com last time I checked, if they
pulled it I can mail you the pdf )
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)
-----END PGP SIGNATURE-----
Main Index |
Thread Index |