I bought the famous $298 eMachines laptop from Wal*Mart.

I shrank the main partition (it has 2 'service' partitions as well) to 50 GB, leaving 150 for Linux, with SystemRescueDisk. I fdisked 2 extended Linux partitions, made ext3 filesystems on them, copied over the system files, but lilo refuses to work, complaining of a fatal raid_setup.

I made a USB Slax (boot) disk. I can boot from it, but that has the same result with lilo. I burned a disk 1 Slack 13.0 CD from the iso image and can mount that, but that doesn't create the setup program.

In my deep dark past (I've been Slacking for 13 years) I always created a boot floppy - but this machine has no floppy drive (my first without!) - and I loaded DOS or Windozze afterwards, something I can't do with Windozze 7. How do I create the same experience with a CD, DVD, or USB drive?

A friend suggested lilo won't work if the Linux partition is on too-high a cylinder ( > 1024 ?). Fixing that would require moving the Windozze partition - I'd rather not.

I'd really like to install Slackware without further tinkering with the partitions or moving Windozzze. But I've forgotten what I'm doing.

Is there a loadlin equivalent anymore? I could put the disk from my old computer in a usb-ata converter and attach that - if that would do any good.

Does grub make this easy? Why does Slackware lack a grub package? Slacking off again?

Thanks for your attention.

I haven't used Slack in a while and lilo probably this century , so ...

The 1024 cylinder thing was a BIOS limitation - nothing recent has that problem anymore. Linux should install (and run) from any partition. I understand grub should be in extras (maybe) on Slack these days.
I'm waiting on a new laptop to be built which will come with Win7 as well - will give me a chance to play with that for the first time. If you haven't solved this by then I'll drop back - might even toss Slack at it to see what Pat has been up to.

This new laptop has 2 'recovery' partitions, 1 about 10 GB, the other about 100MB, that are the first 2 partitions. I made a set of recovery DVDs. Does anyone know if I can delete these 2 partitions, that the recovery DVDs provide the same service?

Thanks for your attention.

RandomTroll

What do you mean by 'copying the system files.' The error then would be that lilo and the kernel would be looking to start from wherever the originals point, not to your current setup. Also, you don't _mount_ the Slack13 CD, you boot to it, logging on and typing 'setup' at the root prompt.

'What do you mean by '"copying the system files."'
I copied the contents of /bin /boot /etc /lib /sbin from my old computer to the new computer and fixed up changed disk references.

'The error then would be that lilo and the kernel would be looking to start from wherever the originals point, not to your current setup.'
I fixed that.

'Also, you don't _mount_ the Slack13 CD, you boot to it, logging on and typing 'setup' at the root prompt.'
Hmmmm... I burnt the Disk 1 iso and it wouldn't boot, but I could mount it. Did I burn it improperly?

I installed grub. I get the grub prompt, from which I can boot into Windozze but when I boot into Linux at the point that it tries to load USB disks (I have none - I disconnected all my USB devices) it reboots.

Thanks for your attention.

RandomTroll

Hmmmmm... I made a mistake. I *can* boot the Slackware disk 1 I burnt. setup crashed the first time it tried to configure the system, but worked the second.

But when I boot it halts just after executing /etc/rc.d/rc.local. Shift Lock keys toggle their LEDs but nothing else responds.

Is there anything _in_ rc.local?

You might try booting to the cd, mounting your hard drive, and then poking around the logs for clues to the hang-up.

If you got as far as rc.local and the machine is connected to a network, you could ssh into it when it hangs and poke around to see what's what. I have a machine that gets stuck there as well (an old desktop) and I have to remotely log in and kill rc.M, after which the login prompt appears. Plus, I have all the other consoles (alt+F1, F2, etc.).

Hi,

The 'md5sum' or 'hash' is very important to learn to use and too regularly get in the habit of utilizing it. You can get a sum checker for M$ if need be; 'md5sum.exe'.

For GNU/Linux the 'man md5sum' will get you all the information to perform the check. You can get the 'man command' from the 'cli' at anytime.

If you downloaded the CD/DVD ISO then be sure to check the md5sum for the original ISO. From the cli;

If the ISO md5 is OK then you should check the image burn with 'CdromMd5sumsAfterBurning''.

This way you will know if the burn was OK!

This will check the download ISO with the known md5sum that you also get with the ISO. You should do the check for any download that you might perform, even a LiveCD. Plus burn the image. Sometimes you may need to lower the burn rate to get a valid burn.


These links and others can be found at 'Slackware-Links'. More than just Slackware® links!

If the install process (finally) went ok, why go backwards to check the download MD5? Wouldn't a corrupted d/l bork the burn?

Hi,

No, garbage can be unfounded or noticed. If the download ISO md5sum/hash is not correct it could be something as simple as a dropped bit that would reflect a change at that point. No telling where. The hash check along with the check for a valid burn would insure a valid install. The install may have finalized but was it valid?

When I dumped my custom-made kernel and used the package kernel I could boot. I've built my own kernel for 12 years. I thought I knew what I was doing.

Thanks for everyone's help.

Well, at least we hounded you until you found it. Man, I've been rolling my own for almost as long and _there_ is a place where every jot and tittle produces a result, and not always what's expected. Happy it was that 'simple,' and good luck figuring out what it was...

And thanks to you, onebuck, your reply was well founded and worth remembering.

Slack on, gents, I'm off to look for a 13_64 live CD to pre-test _my_ future dual-boot beastie.

It seems I need an initrd.tgz now. I hadn't before. Odd.

nonononono!

Kinda.

The kernel's gotten a bit beefy and, as I understand it, needs the ramdisk to throw its massive weight into memory to boot. Do a man for 'mkinitrd' and you'll have the syntax for creating one of your very own... Bare bones requirements are a pointer to the kernel you're using and the filesystem you use to boot (plus the '-c' script to keep things tidy). Once you've created it, run 'lilo -v' and you should be good to go. If you poke around the lilo.conf, you'll see how the bootloader points your choice to the initrd .tgz which was created when you 'mkinitrd'-ed.

Are we still talking about the out-of-the-box kernel? Another problem I've had is with roll-your-owns and mkinitrd. I think its time (for me, anyway) to get off my lazy kiester and copy the kernel directory and make my kernels with pointers to specific directories which won't interfere with standard install structure.

Never thought I'd mention ending laziness when referring to Slack. Sounds oxymoronic.

Get to work,
Geoff.

Quoth wheelsg: 'The kernel's gotten a bit beefy and, as I understand it, needs the ramdisk to throw its massive weight into memory to boot.'

That's not it: I can load the 'huge' kernel, 50 MB, without an initrd.tgz. It's for modules needed to load the kernel. And I was wrong: that's not my problem. I thought it was because when I booted with the non-huge kernel I had the same symptom. Something else is going on.