[linux-lvm] 2.4.20-pre4 and lvm-1.0.5

[Donald Thompson]

On Mon, 26 Aug 2002, Heinz J . Mauelshagen wrote:

> Did the "cd LVM/PATCHES;make;cd $KERNEL_DIR;patch -p1 TheLVMPatch" run
> show any problems?

None, it patched clean. I did however (hopefully completely unrelated) get
rejects on two places with the crypto patch, which I went back and
manually added.

> Do you say that pvscan etc. even fail when /lib/modules/2.4.20-pre4
> is in the module path and *no* modules get loaded at all?

That is correct. I boot, vgscan fails (I'm assuming it segfaults but I
really can't tell during the normal boot process what actually happened)
to detect my VGs, my system goes into a maintenance mode situation since
my /usr is on an LV, I login and run lsmod and absolutely no modules
appear in the list. I try vgscan and/or pvscan and they fail with a
segmentation fault. So yeah, if no module tree exists system boots fine,
if module tree exists vgscan fails regardless of whether modules are
loaded. Yes it defies logic, but I swear thats how it worked:)

> Did you ever retry with a clean build
> "make clean&&make dep&&make bzlilo&&make modules&&make modules_install"
> and an error free depmod run?

I did get a depmod error with sis.o I think it was, which is a video
driver module. I removed that module and depmod ran without errors.

What I did do the next day is with a straight 2.4.19 kernel, patched in
LVM along with the same pptp patch and crypto patch, and the system came
up fine. Two big differences here are obviously I didn't patch in the pre4
patch, but also I didn't use the linux-2.4.19-VFS-lock.patch with the pre4
kernel, whereis with the straight 2.4.19 kernel I did. Just spaced
patching the VFS-lock patch, but I don't really need it anyways.

Either pre4 and LVM don't like one another, or I screwed something up and
I missed it when I was working with the 2.4.20-pre4 kernel. Since I
haven't done a very 'scientific' analysis of where stuff is going wrong,
I'd put my money on operator error at this point.

-Don

> Regards,
> Heinz    -- The LVM Guy --
>
>
> On Sun, Aug 25, 2002 at 02:42:31AM -0600, Donald Thompson wrote:
> > I built 2.4.20-pre4 tonight and patched in lvm-1.0.5. My volume groups
> > fail to showup after and only after I built the modules. In other words
> > some module i've built is screwing up LVM. vgscan gives me a segmentation
> > fault.
> >
> > If I move /lib/modules/2.4.20-pre4 out of the way so that no modules are
> > ever loaded at bootup, everything works fine with my volume groups. As
> > soon as I boot into the kernel with modules available I get segfaults
> > on both vgscan and pvscan. I haven't taken the time to track down exactly
> > which module it might be since i've got around 180 modules built, and
> > honestly I have no clue how to make even educated guess on it. I first
> > thought it was rtc.o since this was the only thing that was getting loaded
> > when the volume groups failed to show. However, I moved rtc.o out of the
> > modules directory, rebooted, vgscan still fails to run, and lsmod lists no
> > modules loaded.
> >
> > Prior to this I was running 2.4.19-pre10 with lvm 1.0.4 patch. I
> > discovered I was still using lvm 1.0.1 utilities though up until the time
> > I built 2.4.20-pre4. I also swapped out the motherboard and CPU on the
> > system prior to rebuilding it, so the configuration changed somewhat.
> >
> > I can boot back into 2.4.19-pre10 while still using lvm-1.0.5 utilities
> > and everything is fine.
> >
> > Anyone have any clues where I might have gone wrong?:)
> >
> > This is a debian sid system. Kernel was compiled with gcc-3.0.4. Binutils
> > 2.11.2. Hardware is: athlon K7 950mhz CPU, Via KX133 type motherboard.
> > Along with the lvm patch that I created for this kernel in the lvm PATCHES
> > directory, I've also patched in a pptp patch (believe it only effects pppd
> > stuff) and the crypto patch 2.4.3.1.
> >
> > TIA for any help.
> >
> > -Don
> >
> >
> > _______________________________________________
> > linux-lvm mailing list
> > linux-lvm at sistina.com
> > http://lists.sistina.com/mailman/listinfo/linux-lvm
> > read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
>
> *** Software bugs are stupid.
>     Nevertheless it needs not so stupid people to solve them ***
>
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
>
> Heinz Mauelshagen                                 Sistina Software Inc.
> Senior Consultant/Developer                       Am Sonnenhang 11
>                                                   56242 Marienrachdorf
>                                                   Germany
> Mauelshagen at Sistina.com                           +49 2626 141200
>                                                        FAX 924446
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm at sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
>