lilo boot problems

lilo boot problems

[maarten]

Hi list,

I've got bitten by the recent bug(?) that lilo seems to have with 2.6.x 
kernels where, upon install, it prints:

Fatal: Trying to map files from unnamed device 0x0000

I've tried everything I could think of in lilo.conf to no avail.  Using google 
there are several (but not many) threads about this specific error but I've 
not seen a solution anywhere yet.

A "raid-extra-boot" line doesn't help any, neither do any other options.

At this point, I'm not even sure where the error comes from as even trying to 
install a normal non-raid lilo.conf now fails.

As an aside, this is what I'm trying to accomplish:
I have a SATA boot drive /dev/sda with the root fs being /dev/sda2.  I have a 
degraded raid1 device consisting of /dev/sdb1, /dev/sdc1, dev/sdd1.  
(I realize this is badly chosen; had the partition number been 2 instead of 1 
I probably would've been able to boot after changing the physical order of 
the drives without any changes...  But as it is, I can't.)
I'm trying to swap sdb and sda so it will boot off the raid1 device (or else 
just of any of the 3 simple partitions that form the raid1 device...) so I 
can then add the former /dev/sda to the array to complete it. 

Any insights...?

Maarten

-- 
When I answered where I wanted to go today, they just hung up -- Unknown

Re: lilo boot problems

[Laurent CARON]

maarten wrote:

>Hi list,
>
>I've got bitten by the recent bug(?) that lilo seems to have with 2.6.x 
>kernels where, upon install, it prints:
>
>Fatal: Trying to map files from unnamed device 0x0000
>
>I've tried everything I could think of in lilo.conf to no avail.  Using google 
>there are several (but not many) threads about this specific error but I've 
>not seen a solution anywhere yet.
>
>A "raid-extra-boot" line doesn't help any, neither do any other options.
>
>At this point, I'm not even sure where the error comes from as even trying to 
>install a normal non-raid lilo.conf now fails.
>
>As an aside, this is what I'm trying to accomplish:
>I have a SATA boot drive /dev/sda with the root fs being /dev/sda2.  I have a 
>degraded raid1 device consisting of /dev/sdb1, /dev/sdc1, dev/sdd1.  
>(I realize this is badly chosen; had the partition number been 2 instead of 1 
>I probably would've been able to boot after changing the physical order of 
>the drives without any changes...  But as it is, I can't.)
>I'm trying to swap sdb and sda so it will boot off the raid1 device (or else 
>just of any of the 3 simple partitions that form the raid1 device...) so I 
>can then add the former /dev/sda to the array to complete it. 
>
>Any insights...?
>
>Maarten
>
>  
>
you have to reconstruct your array first.

Re: lilo boot problems

[maarten]

On Friday 27 August 2004 15:04, Laurent CARON wrote:
> maarten wrote:
> >Hi list,
> >
> >I've got bitten by the recent bug(?) that lilo seems to have with 2.6.x
> >kernels where, upon install, it prints:
> >
> >Fatal: Trying to map files from unnamed device 0x0000
> >
> >I've tried everything I could think of in lilo.conf to no avail.  Using
> > google there are several (but not many) threads about this specific error
> > but I've not seen a solution anywhere yet.
> >
> >A "raid-extra-boot" line doesn't help any, neither do any other options.
> >
> >At this point, I'm not even sure where the error comes from as even trying
> > to install a normal non-raid lilo.conf now fails.
>
> you have to reconstruct your array first.

Eeh...  You're joking, right ?  
The array holds no data as yet so it is possible, but several things seem too 
odd about this advice.  First, there is nothing wrong with this array, it was 
created by mdadm with all the right options. Second, this being linux (not 
windows) I can hardly believe one has to solve problems by no other means 
than starting all over ("reinstall windows") if the only error is due to the 
bootmanager interacting with the kernel (remember, this error doesn't occur 
with other, earlier, kernels).

Anyway, I chose to work around it by installing grub (and it works fine).

Thanks,
Maarten

-- 
When I answered where I wanted to go today, they just hung up -- Unknown

Re: lilo boot problems

[Laurent CARON]

maarten wrote:

>On Friday 27 August 2004 15:04, Laurent CARON wrote:
>  
>
>>maarten wrote:
>>    
>>
>>>Hi list,
>>>
>>>I've got bitten by the recent bug(?) that lilo seems to have with 2.6.x
>>>kernels where, upon install, it prints:
>>>
>>>Fatal: Trying to map files from unnamed device 0x0000
>>>
>>>I've tried everything I could think of in lilo.conf to no avail.  Using
>>>google there are several (but not many) threads about this specific error
>>>but I've not seen a solution anywhere yet.
>>>
>>>A "raid-extra-boot" line doesn't help any, neither do any other options.
>>>
>>>At this point, I'm not even sure where the error comes from as even trying
>>>to install a normal non-raid lilo.conf now fails.
>>>      
>>>
>>you have to reconstruct your array first.
>>    
>>
>
>Eeh...  You're joking, right ?  
>  
>
As you like it.....

I had the same problem, and had to reconstruct the array first (or 
revert back to 2.4.X kernel)

-- 
Soap and education are not as sudden as a massacre, but they are more
deadly in the long run.
		-- Mark Twain

Re: lilo boot problems

[Louis-David Mitterrand]

On Sat, Aug 28, 2004 at 12:44:55AM +0200, maarten wrote:
> On Friday 27 August 2004 15:04, Laurent CARON wrote:
> > maarten wrote:
> > >Hi list,
> > >
> > >I've got bitten by the recent bug(?) that lilo seems to have with 2.6.x
> > >kernels where, upon install, it prints:
> > >
> > >Fatal: Trying to map files from unnamed device 0x0000
> > >
> > >I've tried everything I could think of in lilo.conf to no avail.  Using
> > > google there are several (but not many) threads about this specific error
> > > but I've not seen a solution anywhere yet.
> > >
> > >A "raid-extra-boot" line doesn't help any, neither do any other options.
> > >
> > >At this point, I'm not even sure where the error comes from as even trying
> > > to install a normal non-raid lilo.conf now fails.
> >
> > you have to reconstruct your array first.
> 
> Eeh...  You're joking, right ?  
> The array holds no data as yet so it is possible, but several things seem too 
> odd about this advice.  First, there is nothing wrong with this array, it was 
> created by mdadm with all the right options. Second, this being linux (not 
> windows) I can hardly believe one has to solve problems by no other means 
> than starting all over ("reinstall windows") if the only error is due to the 
> bootmanager interacting with the kernel (remember, this error doesn't occur 
> with other, earlier, kernels).
> 
> Anyway, I chose to work around it by installing grub (and it works fine).

By "reconstruct" I think Laurent meant "synchronize". In effect,
degraded arrays would produce that error with lilo with certain versions
of kernel 2.6.x

-- 
        "You're cute when you're stupid"

Re: lilo boot problems

[Maarten]

On Monday 30 August 2004 10:26, Louis-David Mitterrand wrote:
> On Sat, Aug 28, 2004 at 12:44:55AM +0200, maarten wrote:
> > On Friday 27 August 2004 15:04, Laurent CARON wrote:
> > > maarten wrote:

>
> > > you have to reconstruct your array first.
> >
> > Eeh...  You're joking, right ?
> > The array holds no data as yet so it is possible, but several things seem
> > too odd about this advice.  First, there is nothing wrong with this
> > array, it was created by mdadm with all the right options. Second, this
> > being linux (not windows) I can hardly believe one has to solve problems
> > by no other means than starting all over ("reinstall windows") if the
> > only error is due to the bootmanager interacting with the kernel
> > (remember, this error doesn't occur with other, earlier, kernels).
> >
> > Anyway, I chose to work around it by installing grub (and it works fine).
>
> By "reconstruct" I think Laurent meant "synchronize". In effect,
> degraded arrays would produce that error with lilo with certain versions
> of kernel 2.6.x

Ah, okay.  This would make sense, yes.  But it creates a catch-22 situation 
because I can't add the missing drive to the array when I have booted off 
that single drive, and I can't boot off the array due to this lilo error.
Not that that can't be fixed...

Maarten