DS5000 memory
Ahnjoan Amous
years using "test" at boot prompt and recieved the following.
>>test
?TFL: 3/mem (2: board 5: too many SBEs: 4194304) [KN03-AA]
?TFL: 3/mem (1: board 7, MBE= 8388608, SBE= 0) [KN03-AA]
?TFL: 3/mem (2: board 10: too many SBEs: 4194304) [KN03-AA]
?TFL: 3/ni/ext-lb (3:(xmt [00000006])) [KN05]
?RTCsi/cntl
?TFL: 3/mem (2: board 5: too many SBEs: 4194304) [KN03-AA]
?TFL: 3/mem (1: board 7, MBE= 8388608, SBE= 0) [KN03-AA]
?TFL: 3/mem (2: board 10: too many SBEs: 4194304) [KN03-AA]
?TFL: 3/ni/ext-lb (3:(xmt [00000006])) [KN05]
?RTC
>>
So I moved the board in slot 5 to slot slot 8, slot 7 to slot 10, and
slot 10 to slot 9. This in an effort to identify the problem memory
modules. I then restarted the DS and ran test again with results that
indicate the problem doesn't follow the memory shift. The following
output is from another issuance of "test".
If anyone has an idea of what these messages mean I would appreciate it.
>>test
?TFL: 3/mem (2: board 4: too many SBEs: 31948) [KN03-AA]
?TFL: 3/ni/ext-lb (3:(xmt [00000006])) [KN05]
?TFL: 3/mem (2: board 4: too many SBEs: 31948) [KN03-AA]
?TFL: 3/ni/ext-lb (3:(xmt [00000006])) [KN05]
Thanks
Ahnjoan
--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-...@muc.de
Henry Bent
Hi,
SBEs are single bit errors. They are the sort of errors that ECC
memory, the kind used in DECstations, is designed to identify and
correct. If, however, there are too many single bit errors it
generally indicates a problem with the memory. The first step is to
thoroughly clean all of the contacts on the memory and the board; I
generally use the somewhat low-tech approach of some rubbing alcohol
and and old toothbrush. You can also try a pencil eraser on the
memory module contacts if they have become oxidized. If that fails I
would suspect that the main board is failing and needs to be replaced.
In my experience this sort of thing is increasingly common with pmax
machines; in the last four years I have had three machines (two /120s
and one /25) fail in essentially unfixable ways.
-Henry Bent
Peter C. Wallace
> Date: Mon, 3 May 2010 00:23:58 -0400
> From: Ahnjoan Amous <ahn...@gmail.com>
> To: port...@NetBSD.org
> Subject: DS5000 memory
I vaguely remember going through this years ago, I think it was solved by
dipping the memory module connector area in 99% isopropyl and inserting and
removing it a couple of times (assuming this is a 200/240/260 with the
pin/socket connectors)
Wait till it drys before powering up!
Peter Wallace
Aaron J. Grier
> I vaguely remember going through this years ago, I think it was solved
> by dipping the memory module connector area in 99% isopropyl and
> inserting and removing it a couple of times (assuming this is a
> 200/240/260 with the pin/socket connectors)
I heard that the memory sockets are only rated for a few insertion
cycles? my 5000/240 has definitely had its share of single-bit errors.
wiggling the connectors seems to postpone it for a while.
--
Aaron J. Grier | "Not your ordinary poofy goof." | agr...@poofygoof.com
Maciej W. Rozycki
> > I vaguely remember going through this years ago, I think it was solved
> > by dipping the memory module connector area in 99% isopropyl and
> > inserting and removing it a couple of times (assuming this is a
> > 200/240/260 with the pin/socket connectors)
>
> I heard that the memory sockets are only rated for a few insertion
> cycles? my 5000/240 has definitely had its share of single-bit errors.
> wiggling the connectors seems to postpone it for a while.
It was the original PMAX (DS3100) that had memory module sockets rated
for the maximum of 25 insertion-removal cycles with the recommended number
limited to 5. That would certainly apply to PMIN (DS2100) systems that
were virtually identical and possibly 3MIN & MAXINE (DS5000/1xx &
DS5000/xx) ones that used the same modules.
I recall no mention of a limit for the different-style sockets of 3MAX/+
(DS5000/2x0) systems and cleaning the contacts always fixed any bit error
problems for me. Although I recall I had to be determined at least once,
where I had to go through several remove-clean-insert iterations before
the errors eventually went away. It could have been oxidation that
happened or suchlike.
You can select individual modules to test BTW, to save time that is
considerable with a decently populated system -- the help message for the
memory test will tell you how (I don't remember off the head, sorry).
Maciej