Another discouraging message (was: Re: arla installation problem?)

[Faheem Mitha]

On 25 Sep 2000, Love wrote:

> Faheem Mitha <faheem at email.unc.edu> writes:
> 
> > Dear Arla-drinkers,
> > 
> > I still have little idea how to get things running. I tried typing
> > startarla and got 
> > 
> > root ~>startarla
> > /usr/arla/bin/xfs.o: /usr/arla/bin/xfs.o: No such file or directory
> > Creating device /dev/xfs0
> > 
> > I see that I have /usr/arla/bin/xfs.smp.o and /usr/arla/bin/xfs.up.o but
> > no /usr/arla/bin/xfs.o. However, I do have /lib/modules/2.2.10/fs/xfs.o
> > and /lib/modules/2.2.10smp/fs/xfs.o. (I should add that this is how the
> > spec file in the distribution was set up originally). I am out of my depth
> > here. Is this a problem (sure looks like it) and if so, what is the best
> > way to fix it?
> > 
> > I see that startarla has the lines
> 
> There is not xfs.o since the rpm does have a redhat:ism in itself, it build
> two kernelmodules, one up and one smp, by setting __BOOT_KERNEL_SMP and
> __BOOT_KERNEL_UP, overrideing /boot/kernel.h.

When I changed startarla to be insmod xfs, /lib/modules/2.2.10/fs/xfs.o
gets loaded (that is what the system tells me). From what I understand of   
the spec file, this is identical to /usr/arla/bin/xfs.up.o. Am I missing
something here? I must admit I am not sure why is was necessary to have
two copies of the module in different places.
    
> Since I haven't seen a SuSE computer (except one 2 1/2 years ago, and that
> was a german installation, so that doesn't count :) I don't know what black
> magic is required to make SuSE build a smp and a up (uniprocessor) kernel,
> don't even know if that is possible.

Gee, I hope black magic is not really required, as I don't know any black
magicians. :-) It seemed to compile them OK, as far as I could see.

> startarla doesn't have that magic, but the sysv init script arla.init does.

What magic is that?

I am sorry, but I don't thinking I am really following your comments
above. Are you saying that there is a problem with the kernel module? It
seems to be functioning (sort of). At least, as I said in another message
that I sent out, I seem to be contacting the remote server in test mode
(using arlad -t).

Does this narrow down what might be wrong?

Sorry about cluttering up the list with all this stuff.

                        Best regards, Faheem Mitha.