[Fwd: [Bug 154763] Grub is inadequate without Lilo for backup]

[Mike Bird]

Alan,

On Wed, 2005-04-20 at 06:55, Alan Cox wrote:
> Its unmaintained

False wrt Lilo in general.  Circular argument wrt RH.  Irrelevant since
the long-unmaintained RH version of Lilo just works.

> Its obsolete

False.  Might become true in future if Grub 2 becomes as reliable as
Lilo, which currently seems unlikely with all the extra widgets being
grafted on.

> Nobody in Red Hat tests/qa's it

Circular argument.  Irrelevant since Lilo just works.  Irrelevant,
judging by the effort RH has spent over the years trying to get Grub up
to snuff.

> We can use the disk space for real packages

Falsely assumes Lilo is not a real package.  Contradicts previous RH
statement that disk space is not an issue wrt Lilo.  If disk space is an
issue then drop Lilo's docs which aren't needed during Anaconda.

> Its more productive to fix grub than import prehistoric baggage like lilo

Falsely assumes Lilo is prehistoric baggage.  Falsely assumes Lilo takes
significant effort to import - just leave it in the Core.  False on its
face, as RH has worked to fix Grub for years whereas even the ancient
version of Lilo that Redhat shipped just works.

> lilo doesn't understand file system layouts

Irrelevant - neither does vim or the TCP stack.  Lilo boots reliably. 
That's all we want it to do.

How does an alternative reliable boot loader hurt anything other than RH
egos, particularly if hidden from newbies?

How many times has Redhat helped a RHEL subscriber because Grub changed
its concept of which device was which, or couldn't handle MD, or broke
after a restore, or was used on uncertified hardware, or just
mysteriously broke?

Will Redhat drop Lilo from RHEL too, or is this just a measure to
cripple Fedora?

--Mike Bird