2.1.x and IBM Netfinity
Stepan, Jamison
We have linux installed on an IBM Netfinity 5500 and under 2.0.x everything
works well. When I tried upgrading is when we ran into problems. Here's what
I can say:
The machine has some proprietary stuff in it (ServeRAID and a weird Net
card) so we had to provide our own hardware for some things (Adaptec 2940
card and 3Com 3c900 card). The machine also has two processors (2 400 P2s).
Under 2.0.x only one is recognized since IBM uses the v1.4 BIOS (hence one
big reason we want to move to 2.1).
When I compiled and installed the new kernel the machine reboots and sees
both processors (yay!) and starts working normally. After about 17 to 19
minutes though it completely locks and reboots. It's a hard enough reboot
that the machines built in error light comes on which is only supposed to
happen in case of hardware failure. The machine will cycle and through and
continue this process every 17 to 19 minutes. The first thing I tried was
pulling the second processor and making it run on just one. It didn't change
anything.
I get no error messages, nothing logged to the syslogd, I even tried cat'ing
/proc/kmsg and watching through a crash and nothing happened. The only odd
thing I can see is that there are two lights for the processors on the front
of the machine. Only one is ever solid, the other blinks as it is accessed,
except during a crash, at which point both remain solid. Obviously if I have
only one processor in the machine it's always a solid light, so I don't know
if this means anything.
I am not sure what to do to get some debugging info for people so any help
would be appreciated. I will also try and answer any questions I can to try
and figure out what is going on.
Thanks,
Jamison Stepan
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jamison Stepan
Internet Application Developer
Minnesota Department of Children,
Families and Learning
Jamison...@state.mn.us
jst...@scc.net
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Alan Cox
> Under 2.0.x only one is recognized since IBM uses the v1.4 BIOS (hence one
> big reason we want to move to 2.1).
Umm 2.0.x recongizes MP1.4, it just doesnt recognize MP1.1/1.4 in
some places its allowed (EBDA from memory).
Alan
Doug Ledford
> Hello,
>
> We have linux installed on an IBM Netfinity 5500 and under 2.0.x everything
> works well. When I tried upgrading is when we ran into problems. Here's what
> I can say:
>
> The machine has some proprietary stuff in it (ServeRAID and a weird Net
> card)
Use the PCNet32 nic driver.
> Under 2.0.x only one is recognized since IBM uses the v1.4 BIOS (hence one
> big reason we want to move to 2.1).
Yep, know this one.
> When I compiled and installed the new kernel the machine reboots and sees
> both processors (yay!) and starts working normally. After about 17 to 19
> minutes though it completely locks and reboots. It's a hard enough reboot
> that the machines built in error light comes on which is only supposed to
> happen in case of hardware failure. The machine will cycle and through and
> continue this process every 17 to 19 minutes. The first thing I tried was
> pulling the second processor and making it run on just one. It didn't change
> anything.
Try the following test. Boot with mem=16m and see if it lives any longer.
If that works, then I think we already have at least a clue about the
problem, although not an answer yet.
Stephen may have some other tests he would like you to run. Anyone know
where we can get documentation on the Reliance Champion 2.0LE PCI Chipset
that these machines use? (at least, I haven't seen this exact model of
machine, but I think it uses the Reliance PCI chipset like the 5000 I have
here)
Doug Ledford <dled...@redhat.com>
Opinions expressed are my own, but
they should be everybody's.
Ricardo Galli
> big reason we want to move to 2.1).
I have 2.0.34 and 2.0.35 with SMP 1.1 and 1.4 BIOSes without any problem.
Sure this is not _the_ reason to go 2.1.xxx.
> When I compiled and installed the new kernel the machine reboots and sees
> both processors (yay!) and starts working normally. After about 17 to 19
> minutes though it completely locks and reboots. It's a hard enough reboot
> that the machines built in error light comes on which is only supposed to
> happen in case of hardware failure. The machine will cycle and through and
> continue this process every 17 to 19 minutes. The first thing I tried was
> pulling the second processor and making it run on just one. It didn't change
> anything.
I saw a similar problem in one of our machines when the total memory was
specified in lilo.conf (is this your case?), although the live cycle was a
couple of hours (it depends on available ram and workload). Specifying
(1) MB less in lilo.conf solved the problem (related to bios shadowing, some
bioses can not be changed not to shadow).
If you are going to use SMP and 2.0.xx (I found 2.0.35 is really bad for
SMP,
2.0.33 and 2.0.34 performs a little better/longer), I strongly recommend you
2.0.36 (is en pre19 at the moment but it is rock solid in my machine). It
is a big step from 2.0.35, recognizes all available ram also has _much_
better adaptec drivers.
Good luck.
--ricardo
PS: Did anyone see the report of the contract between TurboLinux and Adaptec
to develop "better Adaptec" drivers for Linux? What does it mean? Are they
going to improve Doug drivers or start from scracht?
Alan Cox
> to develop "better Adaptec" drivers for Linux? What does it mean? Are they
> going to improve Doug drivers or start from scracht?
I was a bit puzzled by this too. However while it may just be they are going to
be shipping the new drivers there are other Adaptec scsi controllers (notably
their weird raid thing, so maybe thats what is involved this time)
Alan
Kurt Garloff
> When I compiled and installed the new kernel the machine reboots and sees
> both processors (yay!) and starts working normally. After about 17 to 19
> minutes though it completely locks and reboots. It's a hard enough reboot
> that the machines built in error light comes on which is only supposed to
> happen in case of hardware failure. The machine will cycle and through and
> continue this process every 17 to 19 minutes. The first thing I tried was
> pulling the second processor and making it run on just one. It didn't change
> anything.
I had similar problems (DFI P2XBL/D w/ two PII-350) and I finally disabled
APM in the BIOS. No problems since!
Check it out!
--
Kurt Garloff <K.Ga...@ping.de> (Dortmund, FRG)
PGP key on http://student.physik.uni-dortmund.de/homepages/garloff
Unix IS user friendly - it's just selective about who its friends are!
Stepan, Jamison
recoginze it no matter how I compile the kernel. Granted I haven't tried 36,
only 34 and 35.
Jamison
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jamison Stepan
Internet Application Developer
Minnesota Department of Children,
Families and Learning
Jamison...@state.mn.us
jst...@scc.net
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> -----Original Message-----
> From: al...@lxorguk.ukuu.org.uk [SMTP:al...@lxorguk.ukuu.org.uk]
> Sent: Monday, November 09, 1998 2:36 PM
> To: Jamison...@state.mn.us
> Cc: linux-...@vger.rutgers.edu
> Subject: Re: 2.1.x and IBM Netfinity
>
> > card and 3Com 3c900 card). The machine also has two processors (2 400
> P2s).
> > Under 2.0.x only one is recognized since IBM uses the v1.4 BIOS (hence
> one
> > big reason we want to move to 2.1).
>
> Umm 2.0.x recongizes MP1.4, it just doesnt recognize MP1.1/1.4 in
> some places its allowed (EBDA from memory).
>
> Alan
-