Parted Magic and syslinux

[krige]

In the following instructions:
http://partedmagic.com/documentation...e-liveusb.html
it is stated that Ubuntu comes with an old version of syslinux which doesn't work with Parted Magic, so it is suggested to install the newest version available at the follow address:
http://www.kernel.org/pub/linux/utils/boot/syslinux/

How do I check the version of syslinux installed on my system? Why Ubuntu comes with an outdated version of syslinux? How do I install the newest version?

[cogier]

Looks to me like you want to create a bootable Ubuntu on USB stick. If this is the case download a Ubuntu ISO image from http://www.ubuntu.com/getubuntu/download. Then go to System>Administration>USB Startup Disk Creator (Ubuntu 9.10) and follow the instructions. You will need a USB stick with at least 2GB of memory.

I don't suggest you download software other than through the Ubuntu Software Centre

[krige]

Thank you cogier but what I wanted is to have an application like GParted or Parted Magic to boot from a USB drive.

I tried using Startup Disk Creator and UNetbootin but none of them worked.

[cogier]

GParted will not run without Linux. You could use one of the small Linux distros.

[krige]

I was using gparted-live-0.5.2-1.iso available at:
http://sourceforge.net/projects/gpar...d-live-stable/
I thought that had integrated some small Linux distro. What would be the use of the iso otherwise?

[oldfred]

I just copied a bunch of ISO to my USB and can boot them. (They have to be set up to allow it)

MultiBoot USB with Grub2 (boot directly from iso files)
http://www.panticz.de/MultiBootUSB

[krige]

I just copied a bunch of ISO to my USB and can boot them. (They have to be set up to allow it)

MultiBoot USB with Grub2 (boot directly from iso files)
http://www.panticz.de/MultiBootUSB

Wow, that would be great if I could make it work!

I meticulously followed the instructions I have found on the site you mentioned, but still can't boot from USB.

I am start thinking there may be something wrong with the USB drive. I have opened it with GParted and have noticed it has the "boot" flag only: should any other flag (lba, lvm, etc) be selected?

Also, "fdisk -l" gives me about it the following information:

   Device Boot      Start         End      Blocks   Id  System
/dev/sde1   *           1        1023     3964094    b  W95 FAT32
Partition 1 has different physical/logical endings:
     phys=(1023, 124, 62) logical=(1022, 124, 62)

Is it normal for it to have different physical/logical endings?

[cheffabe]

Is it normal for it to have different physical/logical endings?

I think it is nothing to worry about. My stick formated by Windows shows a similar result. Deleting the partition and setting a new one with gparted (Option: Round to cylinders) fixed that for me.

[cheffabe]

How do I check the version of syslinux installed on my system? Why Ubuntu comes with an outdated version of syslinux? How do I install the newest version?

Good questions! I was wondering the same, since it is already updated in debian/testing. There's already a bug about that in launchpad but I don't think there will be a fix soon. I think the reason why is because there a patches used with the ubuntu version.

I walked the same path as you did: From pmagic to syslinux. I didn't want to use the syslinux.deb from the debian repros, so I tried to compile it locally. What can I say? I failed since it is more than just a "make install". They use nasm and I'm stuck there.

How far did you get with the Grub-Mulitboot? Did you get the grub boot menu? What was the result of

grub-install --no-floppy --root-directory=<usb-mount> /dev/sdX

?

[srs5694]

Is it normal for it to have different physical/logical endings?

The Master Boot Record (MBR) partitioning scheme attempts to store partition start and end points in two ways:

• Using a cylinder/head/sector (CHS) triplet, which maxes out at about 8GiB
• Using a logical block addressing (LBA) scheme, which maxes out at 2TiB

The message you note says that these two values don't match. This is most likely caused by either a bug in the original partitioning software or a different interpretation of the disk's CHS geometry (that is, how many cylinders and heads it supports) by the original partitioning software vs. fdisk.

The danger here is that two different OSes or low-level utilities will use different values (one using CHS and the other LBA), or that they'll both use CHS but interpret the CHS geometry differently. The result will be inconsistencies in the start and/or end points of the partition(s) on the disk, which in turn will lead to problems. Most likely the problems would manifest as an inability to mount the partition in one OS, but there could be more serious issues. For instance, an attempt to fix what appears to be a corrupt filesystem because of varying interpretations could end up harming it.

That said, my impression is that most modern OSes favor the LBA values; the CHS values are present mostly for historical purposes. Certainly the CHS values are useless on any disk over 8GiB in size. To completely eliminate the chance of misinterpretation, you could convert the disk to use the GUID Partition Table (GPT) instead of MBR; GPT doesn't use CHS values at all, only LBA values.