[plug] re: kernel -smp conflict

[Jon Miller]

Yes same server and it still freezes after a day or two with no error messages.  IBM asked me to apply the latest BIOS and the latest drivers which is where the conflict occurred.  I'm applying the BIOS now and the latest driver.

Jon

Jon L. Miller, MCNE, CNS
Director/Sr Systems Consultant
MMT Networks Pty Ltd
http://www.mmtnetworks.com.au

"I don't know the key to success, but the key to failure
 is trying to please everybody." -Bill Cosby



>>> dsbrown at cyllene.uwa.edu.au 10:05:30 PM 22/07/2003 >>>
Hello, Jon.

On 22 Jul 2003, Jon Miller wrote:

> In conjuction to the previous question, there is one question I've been
> meaning to ask. Since I have both kernels (kernel-smp and standard
> kernel) which should I be loading (smp or non-smp). Also since I made
> the rpm while the kernel-smp was loaded can I assume that if I load only
> the standard kernel and create the rpm again that maybe it'll create a
> rpm for the regular kernel?  Just grasping at straws here.
> 

Well, Andrew has already replied to the question of which kernel to use
for what purpose.   I'm probably the least experienced kernel-compiler on
this list but from what I have seen and done the modules always seem to
need recompiling to suit the specific kernel flavour.   As you have seen,
there are directories for both kernel flavours...
/lib/modules/2.4.18-14/blah, blah, blah  and
/lib/modules/2.4.18-14smp/blah, blah, blah.

That should keep things segregated and if, like me (and Andrew's
suggestion) you were compiling from source rather than doing rpm's, that
should be all you need to do.   BTW the module compilation, because it
takes place in the same phase as the kernel compile itself, doesn't care
what flavour of kernel you're presently running.   Not sure about rpm's -
the game plan could be different but I'd be surprised somehow since it
would imply a chicken-and-egg situation; you need to be running the right
flavour (complete with modules) before you can compile the modules to
suit!

This sounds suspiciously like your favourite x335 server, Jon!   Has it
died recently or did the Broadcom 6.0.2 driver version do a better job?
The 6.2.11 claims (and seems to) fix the occasional null-MAC-address issue
for me on a 205.

Anyway, over to more experienced minds.   HTH,
Denis


> Thanks > > On Tue, 2003-07-22 at 17:29, Jon
Miller wrote:
> > I have a server with a dual Xeon CPU.  I'm using both the main kernel
> > and kernel-smp.  The nic I'm using is having a conflict with the
> > kernel-smp when I try to upgrade the nic. I can see that in the initial
> > install RH installed the drivers in both the kernel and the kernel-smp.
> > In the upgrade of the drivers there does not exist any means to add the
> > updated driver in the
> > /lib/modules/2.4.18-14/kernel/drivers/addon/bcm5700 path. 
> > Do I have to do this manually or is there a way to tell rpm that the
> > files are to exists in both kernel paths?
> > listing of the updated bcm5700 drivers:
> > 
<snip>