Slashdot Mirror
LILO Bootloader Development To End
An anonymous reader writes: For any longtime Linux users, you probably remember the LILO bootloader from Linux distributions of many years ago. This bootloader has been in development since the 90's but development is finally ending. A homepage message reads, "I plan to finish development of LILO at 12/2015 because of some limitations (e.g. with BTFS, GPT, RAID). If someone want to develop this nice software further, please let me know ..."
fucked up partition table
RIP thanks for the memories
Well it looks like Patrick will have to make a change.
Time to offend someone
Thanks for maintaining LILO all these years. I certainly do remember LILO loading on my first installations of Linux. I tried to install it on an IBM PS/2 and the biggest challenge was their micro channel architecture. I don't think I was successful at all, but I learned quickly what the LIL... meant.
http://github.com/gbook/nidb
.. of slackware people.
as a slackware user myself, i have no real problem with grub and i use it on a few machines where it's advantageous, but most of my systems still boot with lilo, and i don't see any need to change them around in the near future.
I was using LILO for quite a while recently, as my newest PC has a terrible UEFI system and the Debian installer couldn't configure GRUB by itself. It served me well in the interim.
Yes. However, my nostalgia for an 8-bit with the OS burned into a factory-produced ROM chip is completely natural. Blip-bloop.
Got them moderator blues I blieve I walk out the do', With these mod-points I been gettin', I 'most never post no mo'
The current comments are mostly trolls and brain-dead idiocy. As typical for the new /.
It wasn't until around 1999-2000 (I think) that distros started replacing LILO with GRUB as the default bootloader. GRUB offered many new powerful features that certainly helped its adoption. That is not to say, though, that LILO didn't have benefits as well (and in some circumstances it still does). It's sad to see that such a pinnacle piece of software contributing to Linux's success is going to be discontinued by the project's primary developer. LILO is such an important part of Linux history that it deserves a place is some kind of "hall of fame". But, it's open sourced so maybe -- just maybe -- someone will pick up the project so that it doesn't die. If not then it will be fondly remembered by those of us who were using Linux back in the olden days (1994 was my first install). Even if it's not continued the source code is informative, but the trolls will not understand that and just keep on using whatever their bootloader and praising whatever it is without understanding wtf it actually does and how the boot process actually works.
...only sausage has two!
How many ends does a circle have?
for the maintainers. The bootloader is a not particularly glamorous problem to work on, but it's critical to everyone and because it involves differing interpretations of standards by manufacturers and various OS developers it had to have been a headache.
Of course later projects had the luxury of a clean sheet, hindsight, and more hardware resources, but without a solid bootloader in the early says of Linux, history would have been very different.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
What about a circular sausage?
I used it a very long time ago and countless people have used it before and since. It's far better to have a definite end date rather than just sporadic updates that grow farther apart and less significant, - leaving people to wonder if it's being maintained or not.
Instead of the maintainer feeling the occasional pang of guilt over not doing anything, they can feel good about what was accomplished during the life of the project and move on to the next thing.
No.
I used it like people use locks on their homes, I had a boot floppy with LILO and no MBR on the fixed disk, so the floppy had to be in the computer to boot as it literally booted from LILO installed on a floppy disk, then LILO found the kernel on the fixed disk and the OS loaded. Take the floppy out and the computer doesn't boot.
Obviously this wasn't going to keep out people that knew what they were doing and came prepared with their own bootable media, but it kept out roommates and their friends. I didn't even have to take the floppy disk all of the way out of the computer, simply unmounting it from the spindle and leaving it in the drive was enough. No one at the time gave that a second thought.
Do not look into laser with remaining eye.
I've used it recently in some distro I tried (maybe a Slackware derivative) and I was surprised it still worked well after all these years (it's an old machine, but anyway...).
I can only say thanks... to Werner, John and Joachim. These three guys have been in the forefront of the good battle and millions of people owe them at least a lot of respect.
I wonder whether there are any scenarios where LILO works and Grub won't...
LILO - Grub, Xfree86 - Xorg, SysV - Systemd, Fvwm - KDE & Xfce, OpenOffice.org - Libreoffice, lots of change for me all these years.
To all you which made the world lighter for others, next time you take a sip of your favorite beverage (wine, beer, tea, whatever), please make a toast to yourselves in my name. Please know there's someone somewhere on this Earth that is very grateful to you and thinks you're the best.
PS: All the other folks in software I didn't mention, you're in this one, too. (mc, Gimp, Konqueror, Kwin, Inkscape, Audacity, the kernel, GNU, etc. etc.)
Poettering will soon release a boot loader with systemd because GRUB and LILO are 'too difficult to use'. Instead you'll have to define the boot loader in an XML and then use /usr/sbin/bootloaderctl to load/unload it. However if your boot loader is not "vendor-defined" to on in your distribution, you'll have to manually load it every kernel upgrade.
Custom electronics and digital signage for your business: www.evcircuits.com
Yes, I'm sorry.
#DeleteChrome
Naw, back in the day lilo was the shit.
Python: 'And then suddenly you have a language which says "we're all stuck with whatever the whiniest coder wants".'
Same for me. My computer booted to Windows95, unless you had the mystery floppy in the drive (1 of hundreds scattered about) -- Then it would boot to linux. What did I use this secret installation for? Nothing useful at the time. I just missed the unix machines I had access to in college.
I still use lilo on all 3 of my boxes. I don't need a pre boot "enviroment" or however grub is described, I just want the OS booted and up and running ASAP. If there's a problem with booting I'll just grab a live DVD , I'm not going to wrestle with a load of cryptic bootloader commands to try and solve the problem.
You know, passwords?
this post contain no useful information, no need to mod it down
Do one thing and do it well.
Yes, GRUB is an improvement over LILO. But it still just does the one thing: Loads and boots a selected O/S. It doen't break other systems or apps. In that sense, it's adoption is nothing like systemd.
Have gnu, will travel.
My personal/playpen Linux box at work is Slackware. It uses LILO. Of course.
There is always a place for software that works. Software that does one thing, does exactly what it's supposed to do, and does it well.
...laura
Not really. Grub didn't try to replace su, login, getty, etc etc etc. It confined itself to being a bootloader. It didn't attempt to force changes on the kernel that would prevent LILO from working, so it was also a genuine choice.
Can't simply replace LILO with systemd?
Sad. Grub is too complicated. I like lilo for the simple configuration file. Great power
Apple II is very simple, thus great power.
LIL-
Good for a few things, to be honest. First one of them is configuration - GRUB (since GRUB 2) is a nightmare to configure. LILO is very simple. Second (less common) is installation - LILO installs itself to the MBR (or start of the partition) and does not require access to a system partition for it's configuration (so if something accidentally deletes /etc/lilo.conf LILO still works). I suppose number three is the small size of LILO, and it being slightly faster then GRUB.
I'm starting to think GNU is the problem with "GNU/Linux" these days.
they're a pain in the ass
Yes, now that's it's 2015 things are so much more advanced with grub and we instead have hang with
GRUB loading stage2...
staring at you
First init, and now this? Where will it end?
“He’s not deformed, he’s just drunk!”
Entirely this. In my opinion grub 2 is where they really went off the rails. When you have a set of configuration files that configure the set of scripts that generate the _actual_ configuration, something has gone horribly, horribly wrong.
Personally I use extlinux and have been very happy with it. You get the classic "single and really simple configuration file" feel of LILO with the subset of bells and whistles that you actually need.
Look into extlinux, it's what I switched to when grub totally went off the rails.
Not sure if you noticed, but Microsoft has replaced its boot loader a few times now (BCD vs boot.ini vs Whatever they added in Windows 8)
I'm starting to think GNU is the problem with "GNU/Linux" these days.
I wonder how long it is til systemd decides that the linux kernel is too confusing for people who just use x86-64 and replaces the kernel.
If someone is prompted with a password prompt they may be inclined to try guessing for awhile. If they're prompted with a disk-boot error they simply turn off the computer and walk away.
Do not look into laser with remaining eye.
Entirely this. In my opinion grub 2 is where they really went off the rails. When you have a set of configuration files that configure the set of scripts that generate the _actual_ configuration, something has gone horribly, horribly wrong.
Those scripts and configuration files sprang into existence well before GRUB 0.97 even. In GRUB 2, they're still not required and the configuration file is not hard to hand-write if you don't want the auto-OS detection and config generation said scripts provide :P
GRUB doesn't do stages anymore. That would be a "rescue>" prompt. It's a little better, and can usually help to save from minor configuration mishaps, though not useful for more major oopses (maybe you changed file systems and forgot to update GRUB, now the boot.img doesn't understand how to read the file system...). It also gives a lot of pause to the thought of "Oh crap, what are GRUB's commands?"
I only switched to GRUB because of installed defaults and my laziness. LILO was, imo, always simpler, cleaner, made more sense.
GRUB works, but it feels like it's been engineered to be way more than it needs to be, and, in the process, it starts to suck. As an example, I started looking into GRUB theming (hey, a pretty boot screen would be nice). Turns out I could never convince GRUB to use a TTF font and display the table (and items in the table) correctly. That's a feature that may as well not be in there -- if it doesn't work, turf it. I didn't *need* it, but I spent a bit of time trying to get it to work and failing, which I wouldn't have bothered with if it just weren't there. Or there's the interactive boot -- works if you already know by rote all of the GRUB internals, at the currently installed level; completely useless otherwise. That being said, it does still work, and I still use it to select an OS -- I just miss LILO.
Also, LILO was quite explicit about the "you need to run me to update the MBR" ruling. Which I found kind of comforting: if you didn't make the active choice to update your MBR, you didn't make the active choice to break it with a bad config.
Also, diversity leads to better overall design across the spectrum. Any time a competitor is lost, it's sad for the ecosystem as a whole.
Stating that LILO and GRUB were "confusing and broken", Lennart Poettering has announced that SystemD will take over MBR management. "It's just a small step towards complete system domination", the great leader was heard to muse, "After all, PulseAudio did so well and everyone is loving how much easier SystemD is than init scripts. What could possibly go wrong?".