I need to add a partition to my drive to install mepis, but qtparted* gives me an error when i select the drive "sdb". I can view "sda" fine, but it crashes on the sdb. I have 7 partitions on the drive currently with 3 distros running perfect. I dont need mepis, but i'd like to try it. I am having a problem with the partition utility in ubuntu where it shows me that the entire drive is "unallocated" space. These are both gnome based distros and i have only kde on the other 3 distros. Am i over the limit with partitions, and why do the other ones work flawless?

If you are not aware of the basic rules of partitioning, like

(1) There must be a maximum of 4 primaries

(2) If a logical is formed then an extended partition will be created at the expense of using up one primary parition.

(3) All logical partitions must be in a continuous chain and cannot be broken into two haves

(4) Whenever the partition table has been change the PC should be rebooted or system is asked to update the partition table as the kernel could be still using the old one.

(5) There can only be one extended partition.

you could have forced a partitioning tool to modify the partition table in such a way that it becomes confusing and consequently regarded as corrupted. In such a case some partitioning tools will play safe and treat your hard disk as empty with unallocated space, even though you may have 30 OS inside all bootable and work happily inside. It is all to do with a partition table that a partition tool does not know how to cope.

You will find there is a packing order in the partitioning tools. Those with GUI are the first ones that give up.

The next one you can try is "cfdisk" in terminal mode.

The last one that you could use when all other partitioning software has abondoned you is "fdisk"

Get a second opinion of the health of your disks from cfdisk and fdisk.

If you log in as root just type
or
or

Here is the table from the ubuntu livecd.

Disk /dev/sda: 120.0 GB, 120034123776 bytes
255 heads, 63 sectors/track, 14593 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 * 1 14592 117210208+ 7 HPFS/NTFS

Disk /dev/sdb: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors/track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/sdb1 * 1 25496 204796588+ 7 HPFS/NTFS
/dev/sdb2 25497 29981 36025762+ 5 Extended
/dev/sdb3 28109 29153 8393962+ 83 Linux
/dev/sdb4 29154 29535 3068415 83 Linux
/dev/sdb5 25497 25687 1534176 82 Linux swap / Solaris
/dev/sdb6 25688 27216 12281661 83 Linux
/dev/sdb7 27217 28108 7164958+ 83 Linux
/dev/sdb8 29536 29981 3582463+ 83 Linux


opensuse - sdb3
ubuntu - sdb4
swap - sdb5
/home - sdb6
mandriva 2007 - sdb7
pclinuxos - sdb8

Yep your partition table is illegal and broke the Rule (3) in Post #2

hda1 to hda4 are reservred names for the 4 primaries and logical partition always start at the 5th position.

As the Rule (2) says if you want an extended partition it can only come from a primary and in this case it is sdb2 with its boundary defined between cylinder 25497 to 29981.

All logical partition must be inside the entended partition in a continuous chain.

You must have over-ruled Linux by changing the sdb7 and sdb8 into sdb3 and sdb4 respectively and that is the confusion because sdb3 and sdb4 are primary partitions and cannot be inside an extended partition. Later addition automatically used up sdb7 and sdb8 position. Your sdb3 and sdb4 have cylinders sandwiched between sdb7 and sdb8. That is a crime in Linux.

Your are going to have to do some work here to bring it back to normal.

You can try other alternatives but to me the only way out is to change sdb3 and sdb4 back to the sdb8 and sdb9 position, therby automatically put the old sdb8 to the new sdb10. The new set up should have no sdb3 and sdb4.

Thus you need to use a Live CD or just use Mandriva in sdb7 to do the following

(1) Make a record the partition sizes, in both Mb, sectors and cylinders.

(2) Use fdisk to delete partition sdb3 and sdb4.

(3) You then exit fdisk and use cfdisk to carry out the partitioning to re-created two logical partitions in "exactly" the same positions of the deleted sdb3 and sdb4. You will find cfdisk will name them automatically sdb8 and sdb9 and your previous sdb8 will automatically become sdb10.

(4) Yout partition table will now be recognised by every partition tool and regarded as healthy but your Opensuse, Ubuntu and PClinuxOS would not boot because their partition references are all wrong.

(5) You should reboot at this point and afterward mount sdb8,9 & 10 to amend the partition refereences in in Grub's /boot/grub/menu.lst ( for Suse and Ubuntu) and Lilo's /etc/lilo.conf (for PCLinuxOS) and the /etc/fstab in all 3 distros.

(6) Don't know how you boot them so can't advise further but your distros should now be all bootable.

Don't lose heart on the above. You will have to do the same if you want to move the Linux around in different partitions. So treat it as an exercise to learn how to do such thing.

Other may ask you to re-format and re-install everything but I see absolutely no need. I shuffle my Linux this way all the time myself.

If i delete sdb3 and 4 and create new partitions of similar size, will it name them in the correct order? I can reinstall opensuse as it is a fresh install anyway, and i'd do it last as it recognizes the existing distros perfectly (other bootloaders dont work as good).

You can learn a lot from the post #4

The whole point of using the right partitioning is it "will not" allow you to go astray with creation of partition.

You current partition table will not be accepted by cfdisk but fdisk would still read it and so can delete sdb3 and sdb4 for you.

After the deletion I ask you to change to cfdisk because your partition table becomes healthy by then and so cfdisk can operate on it. In it you will find between sdb7 and sdb8 there is a dead space left by your earlier deleted sdb3 & sdb4. If you create 2 logical partitions within the space you can "see" cfdisk automatically shift sdb8 down to eventually sdb10.

I was hoping you can see it yourself and understand it without me explaining it here. There is no better teacher than the box itself.

Your sdb5 to sdb7 are safe and not touched so all of them are good and Mandriva can be used for the rescue work without requiring a Live CD. As you didn't tell us how you boot them so I would suggest you to let Mandriva's Grub to take over the MBR as follow
Thereafter you can boot to Mandriva all the time when Suse, ubuntu and PCLinuxOS are unavailable. You can later edit Mandriva's Grub to boot all the systems.

If you have chosen Lilo as the boot loader then the equivalent command is
to be issued as the root user at terminal.

You can avoid the work by re-installation but equally you can learn a lot of useful Linux features in the process by following Post #4. It is your call. We wouldn't mind if you go either way. Let us know if you still have a problem.

You don't have to worry about the boot loader as that is easiest bit.