I've been following this thread for some time now.
I've seen a lot of opinions and preferences, some
based on fact, but some also based on assumption
and, in a few cases, ignorance.

There is no such thing as "superior" or "better"
in the overwhelming majority of cases, and boot-
loaders are one of them.  GRUB is a dynamic boot-
loader that reads information at boot-time. LILO
is an absolute address bootloader that is con-
figured at install-time. A number of the com-
ments made on either are just technically wrong.

E.g., GRUB's installer is not more/less dangerous
than LILO, and LILO has its firmware (BIOS) order
to Linux device mapping issues as well. GRUB has
label and install options like LILO as well.

The big difference between GRUB and LILO con-
tinues to be boot-time. If the first stage is hosed
in either, you're screwed with both. If GRUB's
2nd stage is inaccessible, you're screwed whereas
LILO will still load. Of course if LILO is misconfigured
or a order, storage or other change occurred,
you're screwed, whereas as long as GRUB's 2nd
stage loads, you can dynamically modify things.

Okay, with that out of the way, when does one
still use LILO - or more importantly - why do the
overwhelming majority of distros still ship LILO,
even though rarely is it the default? Because
GRUB cannot understand every disklabel and/or
filesystem, not even on the PC. E.g., I use LILO
to boot a filesystem located in a LDM Disk Label
(e.g., "dynamic disc"), which I have run into where
it was required.  Is the commonplace? No, it's
the exact opposite.

This whole thing isn't about "better" or "excuses"
or countless other, rather "subjective" comments.
This goes back to my original concept of under-
standing the differences, let alone the reality of
how booting works and what issues could arise.
You could write a book on this, just for the PC -
and, personally, I'd love to shouve it down the
throat of people because 10 times out of 10, they
do *NOT* remotely understand GRUB and LILO
at all. That includes many comments made in
ignorance in this thread over and over on both
sides.

In reality, you've have had to do a lot of non-
standard and, even more so, non-PC and/or non-
PC BIOS-ATA to even begin to understand them.
So, again and in reality, we are testing junior sys-
admins in LPIC-1. We have a finite number of ob-
jectives over a finite number of questions to
cover.  So do any of these greater boot-time
questions fit, including where LILO must be used?
No! As someone like myself, who still uses LILO
(even touched MILO the other day again), let alone
U-boot and the in's/out's of Anvin's SYSLinux
quartet of loaders, says ... it's GRUB for LPIC-1.

You should still mention LILO in the objectives for
completeness, but I would not be including any
LILO-specific questions except where they can
overlap with GRUB and are good, general know-
ledge details.  This discussion has largely missed
that, and as someone who has seen almost ever
PC, non-PC, non-standard bus/storage and
countless other hacks and approaches, I do not
see any, single reasaon for a focus for LILO to
remotely stick at all. I could easily focus on many
other objectives that are better tests for the
best LPIC-1 candidates than someone who knows
LILO.

I have this same argument over knowing assem-
bler, especially people who assume any give pro-
grammer can remotely "optimize" code for
today's superscalar architectures. Learning a 3G
language like C teaches just as much about com-
puter organization. It sickens me when people
argue that assembler teaches "low-level" wheb
that's total BS, and it rarely teaches optimization
these days. I say that as an EE with an educational
specialty on computer archietecure with several
years experience in the semi-conductor industry
(Let alone embedded application).

Same deal on LILO. Proponents are often mis-
guided and often dead-wrong on several concepts
on booting. GRUB is the most popular because the
need for LILO is greatly reduced (virtually
eliminated). Statements on GRUB's alleged issues
are often incorrect or at least red herrings.
LILO should be "known of" and left at that now.
There are other things to focus on, even for boot.

--  
Bryan J Smith - mailto:[EMAIL PROTECTED]  
http://www.linkedin.com/in/bjsmith
Sent via BlackBerry from T-Mobile  
    
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev

Reply via email to