From: David Evans <[email protected]>
> For the record, I didn't say Red Hat shipped LILO.
> I said its documentation still mentions it (though unsupported).
Pardon? The Red Hat documentation certainly does _not_ do this.
Please understand that. Unless you're reading really old, legacy
documentation from earlier last decade (i.e., 2003 is the last [1]).
Before I make my point, let's revisit exactly what was said ...
First off, you stated:
"and mentioned as an alternative to GRUB in the RHEL docs."
Miguel then asked you where, and referred to the RHEL5 docs, showing
no such passage.
I then pointed out as of RHEL5, Red Hat doesn't even ship it as an
alternative. I don't want to nitpick on this point, but please
understand something. Red Hat does _not_ go off suggesting
un-shipped, un-supported software when it comes to the init process.
If it did, I'd _immediately_ file a Bugzilla. I have in the past when
something was still in the documentation but was yanked.
E.g., as of early RHEL5 Updates ("Dot Releases"), Red Hat now
publishes far more detailed, long (i.e., 100s of pages in some cases)
of "Technical Notes" separate from the shorter "Release Notes." This
covers all Bugzillas, rebases, deprecations, removals, etc...
The only time I see Red Hat "mention" something "unsupported" is to
point out that either _shipped_ software is either removed or added as
a "Tech Preview" (to be possibly supported in a future update/point
release). But I don't ever remember Red Hat mentioning something not
shipped and not supported in its documentation -- it would be rare.
So ... back to my point ...
Most specifically, the last time LILO is _ever_ mentioned in RHEL3
[1], which is nearly even ELS termination (End-of-Life was 2010). And
it is shipped and _supported_ as an alternative in RHEL3.
In RHEL4, it is shipped and supported, but doesn't get a mention in
the section [2], other than an appendix reference that refers back to
boot (but with _no_ "lilo" itself actually mentioned). In fact only
elilo is referenced (for IA-64). I need to read through all of the
Release Notes, but it might have been deprecated early on.
I would _never_ suggest Red Hat mentioned unshipped and unsupported
software in its documentation. It's dangerous from a support
standpoint and, again, I would immediately file a Bugzilla if I saw
such..
> There is no doubt that GRUB is the way commercial Linux does things.
Fedora dropped it awhile back as well. As I pointed out, you will
_not_ find it in even Koji.
> However, LILO is not dead. I was only suggesting that entirely removing it
> from LPIC 2
> may be premature. (There's also elilo as mentioned elsewhere on this list.)
That was me as well. Maybe you should read my posts and understand my
points. I wouldn't exactly hold up "elilo" as an example, as it's
very, very modified from LILO (including in troubleshooting). Given
this fact was also pointed out in the current objective by Martin ...
From: "Martin Møller Skarbiniks Pedersen" <[email protected]>
> Then you also have to teach the different "error" message from lilo
> like LI or LIL when booting and lilo fails.
So ... are we going to add the IA-64 aspects of "elilo"? And yes,
I've supported IA-64 systems, some of the larger SGI Altrix units with
both Red Hat Enterprise and SuSE Enterprise solutions.
Understand the IA-64 target is waivering from virtually ... well ...
everyone. IA-64 was developed, with elilo, back in '99, for the Red
Hat Linux 7 (RHL7) release. It has been dropped by most everyone
since 2006. Only Red Hat is continuing to support it through 2017,
2020 with ELS, on RHEL5, per its guaranteed Enterprise Life Cycle --
Novell-SuSE has dropped it much sooner.
In fact, the Intel IA-64 target -- which was clearly led by Red Hat at
the time (SuSE was partnering with AMD on x86-64, which Red Hat
adopted later, and vice-versa) -- along with ANSI C++ enforcement
(resulting in the long-term object compatibility advantage Red Hat has
had since RHL7+/RHEL2+), was the reason why Red Hat -- who was the
official, FSF-GNU maintainer of the new GCC 3 release development (Red
Hat had recently acquired Cygnus) -- finally forked GCC 2.96** from
current GCC 3 developments for RHL7 (and RHAS2/RHEL2). Alan Cox
personally responded to my "open letter" on RHL7 and explained this,
among other things. Several other Cygnus, now Red Hat, employees
explained the whole reason for pushing out GCC 3 early, and IA-64 was
the main reason (along with the "bonus" of "throwing the ANSI C++
switch" with it -- in the same attitude as the prior GLibC2 adoption
in RHL5 too, "it has to happen sooner or later").
> As for teaching it, goodness. Make a buried note in the course material: LILO
> was here.
> Here's what a lilo.conf looked like. You have to run lilo when you make a
> change.
> Install it on your lab machine as a bonus if you like. Next bullet point.
> That covers 213.1.
If you want to add a section for "elilo" on IA-64, I'm all for that.
But I would yank LILO for x86 (IA-32e/x86-64 aka EM64T/AMD64).
There's only so much that can be covered in the objectives.
I.e., There are more marketshare arguments for covering more SMTP and
other services than lilo's marketshare in my view. ;)
-- bjs
**P.S. Red Hat had skipped GCC 2.8, after 2.7, adopting Cygnus EGCS
(Enhanced GNU Compiler System) 1.1 in RHL6 instead, which became the
2.91 series. Red Hat skipped 2.95 for its platforms for various
reasons (even though they developed most of it, they stuck with EGCS
1.1/GCC 2.91), largely because it didn't solve the continue C++ ABI
issue (and had its own C++ object compatibility issues with GCC 2.8
and 2.7, much more EGCS), and the rest is history. The ANSI C++
enforcement aspects of 2.96 (which was also the case with 3.0 and 3.1,
Red Hat also shipped 3.1 in RHL7/RHAS2/RHEL2) broke a lot of software,
but Red Hat got its IA-64 target. A lot of people outside of Red Hat
made a stink of that, largely because they had to make their code ANSI
C++ compliant, but at the time, Cygnus-Red Hat was the authority and
driver in GCC 3 development for many years (until much of the
mindshare went to Code Sourcery by later last decade).
[1]
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/3/html/Reference_Guide/s1-grub-lilo.html
[2]
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/4/html-single/Reference_Guide/index.html#s2-boot-init-shutdown-loader
--
Bryan J Smith - Professional, Technical Annoyance
http://www.linkedin.com/in/bjsmith
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
