FC2 : Dell Inspiron 5150 : Installation problem

Guys,

Need your guidance here. This is driving me nuts.

I'm trying to install FC2 in my Dell Inspiron 5150 laptop, but the installation is
giving some confusing information.

This laptop is now having SuSE with XP. The Linux partitions were formatted
using Partition Magic and I want to get rid of SuSE with FC2.

Partition space : XP - 25GB, SuSE - 12.5GB, Dell Utilities - 2GB+

I've all the 'fdisk' entries, as I expect that FC2 might screwup XP, due to now
well know issue. Anyways...

In my dry run, if I choose

Automatic, it shows

hdc 38154MB IC25N040atcs05-0

Device Mount Point Type Format size(MB) Start End
Hard Drives
/dev/hdc
/dev/hdc1 vfat 31 1 4
/dev/hdc2 ntfs 25988 5 3317
/dev/hdc3 /boot ext3 102 3318 3330<--- newly added
/dev/hdc4 Extended 12033 3331 4864<--- newly added
/dev/hdc5 reiserfs 11013 3331 4734<--- newly added
/dev/hdc6 swap 1020 4735 4864<--- newly added

Manual, it shows

Device Mount Point Type Format size(MB) Start End
Hard Drives
/dev/hdc
/dev/hdc1 vfat 31 1 4
/dev/hdc2 ntfs 25988 5 3317
/dev/hdc3 Extended 12135 3318 4864
/dev/hdc5 reiserfs 11986 3318 4845
/dev/hdc6 swap 149 4846 4864

But choosing /dev/hdc5( or anyone for that sense ), it complains "You have not defined a root partition( / ), which is required for installation of Fedora Core to continue."

What's the problem ?

Information from fdisk -l
fdisk -l
disk /dev/hdc:40.0GB
255 heads, 63 sectors/track, 4864 cylinders
units = cylinders of 16065 * 512 = 8225280 bytes

device boot start end blocks id system
/dev/hdc1 1 4 32098+ de dell utility
/dev/hdc2 * 5 3317 26611672+ 7 HPFS/NTFS
/dev/hdc3 3318 4864 12426277+ f w95 ext'd( LBA )
/dev/hdc5 3318 4845 12273628+ 83 Linux
/dev/hdc6 4846 4864 152586 82 Linux swap

fdisk -l /dev/hdc1
disk /dev/hdc1: 32MB, 32868864 bytes
16 heads, 63 sectors/track, 63 cylinders
units = cylinder of 1008 * 512 = 516096 bytes
device boot start end blocks id system
/dev/hdc1p1 3412043 903260 883057033+ 0 Empty
Partition 1 has different physical/logical beginnings( non-Linux?): Phys=(513, 6, 57) logical=(3412042, 4, 59)
Partition 1 has different physical/logical endings: Phys=(138, 182, 0) logical=(903259, 5, 30)
Partition 1 doesn't end on cylinder boundary.

/dev/hdc1p2 ? 1281596 3036729 884586932+ 72 Unknown
Partition 2 has different physical/logical beginnings( non-Linux?): Phys=(101, 107, 32) logical=(1281595, 7, 5)
Partition 2 has different physical/logical endings: Phys=(370, 114, 47) logical=(3036728, 3, 57)
Partition 2 doesn't end on cylinder boundary.

/dev/hdc1p3 ? 218938 1538565 665092101 69 Unknown
Partition 3 has different physical/logical beginnings( non-Linux?): Phys=(39, 103, 32) logical=(218937, 12, 22)
Partition 3 has different physical/logical endings: Phys=(115, 111, 46) logical=(1538564, 15, 18)
Partition 3 doesn't end on cylinder boundary.

/dev/hdc1p4 ? 1 16645 8388608 20 Unknown
Partition 4 has different physical/logical beginnings( non-Linux? ): Phys=(32, 32, 32) logical=(0, 1, 21)
Partition 4 has different physical/logical endings: Phys=(345, 32, 19) logical=(16644, 2, 21)
Partition 4 doesn't end on cylinder boundary.

Partition table entries aren't in disk order

fdisk -l /dev/hdc2
disk /dev/hdc5: 27.2GB, 272250352640 bytes
units = cylinder of 1008 * 512 = 516096 bytes
device boot start end blocks id system
/dev/hdc2p1 ? 216399 1904881 850995205 72 unknown
Partition 1 doesn't end on cylinder boundary

/dev/hdc2p2 ? 723265 1262922 271987362 74 unknown
Partition 2 doesn't end on cylinder boundary

/dev/hdc2p3 ? 167316 167316 0 65 Novell Netware 386
Partition 3 doesn't end on cylinder boundary

/dev/hdc2p4 ? 2671568 2671619 25817+ 0 Empty
Partition 4 doesn't end on cylinder boundary

Partition table entries doesn't end on cylinder boundary.

fdisk -l /dev/hdc3
Nothing returned

fdisk -l /dev/hdc5
disk /dev/hdc5: 12.5GB, 12568195584 bytes
units = cylinder of 1008 * 512 = 516096 bytes
Disk /dev/hdc5 doesn't contain a valid partition table

fdisk -l /dev/hdc6
disk /dev/hdc6: 156MB, 156248064 bytes
units = cylinder of 1008 * 512 = 516096 bytes
Disk /dev/hdc6 doesn't contain a valid partition table

Thanks.
Kramer.

Most of the fdisk entries are over my head, but I can explain what it is trying to do on the automatic, and maybe then you can accept it, or mimic it better in manual.

First off, it wants to end up with 3 partitions:

/boot - of about 100 MB
/ - of most of the space you give it to work with
swap - of about double your ram or what looks to be 1 GB

Those are its default partitions.

-----------------------------------------------

In automatic, it is leaving your vfat partition of about 32MB alone. This is the Dell diagnostics stuff. I don't see anywhere the 2 GB you say is Dell restore. I think you are remembering another computer there, like an HP or a Compaq. Dell gives you the CDs and does the 32 MB diagnostics partition, not a hidden restore partition.

The next partition is the NTFS 25 GB partition you have Windows on. It looks to me, when you compare the start and end locations with the beginning of your fdisk info (5 and 3317) that is trying to leave the NTFS partition just the way it finds it.

Then it is making a new 100 MB ext3 partition called /boot at hdc3. It will install your boot and grub files there. When you did it manually, you did not define a location for this /boot partition, and it was complaining about that.

The next entry isn't a partition like the others. You can only have 4 primary partitions, and you are ending up needing more than 4 partitions all together, so it is making hdc4 a primary Extended partition and then dividing it into 2 logical partitions at hda5 and hda6. So don't think of hda4 as a partition on its own. It is a "placeholder" for the sum of hda5 and hda6.

It is leaving most of the remaining space as ReiserFS, I guess because it found it that way and it probably only cares that the /boot partition be ext3. So it is making hdc5 as your / partiton and sizing it at about 11 GB and making it ReiserFS.

Finally, it is making a swap partition at hdc6, sized at about 1 GB.

---------------------------

When you did it manually, you had the extended Primary partiton at hdc3 instead of their hdc5. That is ok. Then, there is no hdc4 because logical partitions always start numbered at hdx5, and they skip hdx3 or hdx4 as necessary to start at hdx5. So, you could manually try to make a hda5 for /boot at about 100 MB, and then a very big hda6 for the main / partition and then a smaller (500 MB or 1 GB) hdc7 as the swap partition. But why not just go then with their automatic? It is the same as far as both ways seem to be handling the vfat and NTFS partitions which is really all you care about, I think.

I'm going to boot over into Linux and try to get some fdisk info from my machine to try to make sense of your info. The one entry about 16 heads is something I have seen in the bug reports, so it concerns me, but it might have to do with the way Dell makes the hidden diagnoistics partition, and that in turn is messing up all the info on the subsequent start/end locations, and it may mean "nothing" in the end, but that gets over my head. I had that same partition when I first installed Linux, but along the way I got rid of it with Partition Magic so I no longer have it to compare directly with yours. If I see anything interesting in my fdisk, I'll post here. And then I'm going to be quiet and let other people get a chance in case I am all off base somewhere or everywhere

Kramer,

Here is my fdisk stuff:

[root@LINUX jrittvo]# fdisk -l

Disk /dev/hdc: 60.0 GB, 60011642880 bytes
255 heads, 63 sectors/track, 7296 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hdc1            5887        6273     3108577+  83  Linux
/dev/hdc2   *           1        5886    47279263+   7  HPFS/NTFS
/dev/hdc3            6274        6376      827347+   c  W95 FAT32 (LBA)
/dev/hdc4            6377        7296     7389900    f  W95 Ext'd (LBA)
/dev/hdc5            6377        6389      104391   83  Linux
/dev/hdc6            6390        7198     6498261   83  Linux
/dev/hdc7            7199        7296      787153+  82  Linux swap

Partition table entries are not in disk order

The hdc1 I have is sort of what used to be the 32MB Dell diagnostics partition, that I enlarged at some point with Partition Magic to about 3 GB and then converted it to ext3 with Partition Magic, and I use it as a spill over partition for Fedora, like to have room for building kernels, etc.

hdc2 is my WindowsXP partition, NTFS, about 45 GB.
hdc3 is a 700 MB FAT32 partition I use for transferring stuff between Windows and Linux, since they can both read and write to FAT32
hdc4 is the placeholder for the next 3 logical partitions....
hdc5 is the 100 MB /boot partition
hdc6 is the 6 GB main / partition
hdc7 is the 750 MB swap partition

My partition table entries are out of order just like yours are, and I think that is because we both used Partition Magic along the way and it does some out of order so it can leave the WinXP boot info alone without having to alter the Windows boot.ini file.

Now, when I start doing like you did and getting "fdisk -l /dev/hdc1" and "fdisk -l /dev/hdc2" info, it turns to nonsense:

[root@LINUX jrittvo]# fdisk -l /dev/hdc1

Disk /dev/hdc1: 3183 MB, 3183183360 bytes
255 heads, 63 sectors/track, 387 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/hdc1 doesn't contain a valid partition table

--------------------------

Disk /dev/hdc2: 48.4 GB, 48413965824 bytes
255 heads, 63 sectors/track, 5885 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

     Device Boot      Start         End      Blocks   Id  System
/dev/hdc2p1   ?       13578      119522   850995205   72  Unknown
Partition 1 does not end on cylinder boundary.
/dev/hdc2p2   ?       45382       79243   271987362   74  Unknown
Partition 2 does not end on cylinder boundary.
/dev/hdc2p3   ?       10499       10499           0   65  Novell Netware 386
Partition 3 does not end on cylinder boundary.
/dev/hdc2p4          167628      167631       25817+   0  Empty
Partition 4 does not end on cylinder boundary.

Partition table entries are not in disk order

hdc2 is my NTFS partition, not the 4 individual partitions it lists above, so I think the fdisk -l command on anything other than the main disk itself just gives spurrious results, or at least nothing I can make any sense of. I don't think you can use it on a partition.

hdc2 is my NTFS partition, not the 4 individual partitions it lists above, so I think the fdisk -l command on anything other than the main disk itself just gives spurrious results, or at least nothing I can make any sense of. I don't think you can use it on a partition.

The 4 others are part of hdc2, I think its by cylinders or what ever it is. Notice the names hdc2p[n]

You're absolutely right. I gave wrong info. Dell Utilities have occupied ~32MB and it shows up in Partition Magic.

{ Interesting enough is the missing 1.75 GB as its a 40GB disk and the last time I called Dell, they told that about 2GB is allocated for Dell Util., Now does that mean that they've hidden space apart from the util that's occupied 32MB - totally bizarre. Anyway, that just made me thinking... }

And PM shows the Linux partition as Ext2 and Linux swap only.

The missing 2 GB are what get eaten up in making a MBR, partition tables, formatting, etc, and the different ways people talk about 1 GB vs. 1,000 MB or 1,024 MB and all that. (Never trust what the Dell people on the phone tell you -- they are mostly clueless You'll get much better info on their message board forums.) There have actually been class-action lawsuits filed because of the differences between simple quoted disk sizes in GB and actual formatted capacities as reported in detail. I think my 60 GB formatted as a single NTFS partition with the Dell 32 MB thing gone was like 56 or 57 GB. There is nothing else hidden there though. You just lose the few GB in a "numbers game".

I don't think PM knows much about ReiserFS, so I think it just lumps it together with ext2.

And I still think the fdisk -l results from a partition are not useful. It is saying for instance that part of my NTFS partition is "Unknown" and another part is "Novell Netware 386". That is just wrong information. It says even stranger things when I do it on the other partitions -- filesystems I have never even heard of.

None of this is helping you with whether you will get hit or not with the bug, though . . . I hope someone can chime in on that part.

Well, well, well. FC2 rocks !

Without an hitch it worked. While I was prepared for the worse, FC2 didn't screw up anything. Installation was a breeze, infact so fast, custom installation( of 3.5 GB ) got over in under 20 mins !!!!

Got a bulk load to up2date, but I'm very releived that it didn't give head-aches of fire fighting the partition information. Bye-bye SuSE.

Thanks to all for the info.

jrittvo : thank you very very much. You gave lot of valueable info.

I'm really glad to hear it went smooth. The up2date will take a while because it needs to go through lots of dependencies and sometimes the servers are overloaded it seems, but once you get current, it will only be a few updates every so often from then on. If you are looking at like 200 packages now, you might want to try going after them in small groups till you get a feel for if it is spending time thinking or if it is spending time stuck and needs to be shutdown and restarted. It will get stuck from time to time.

you're 100% correct. I've about 130+ to be updated. I've already updated the kernel.

But one weird thing - if I choose 15-20 packages, it hangs no matter what, so I kill it and then I choose the 1st 5 or so, it works poooh...

I think it gets the package headers when you select files, and then it starts going through the dependencies for each one, one at a time. Like Package A needs Package B but will break Package C..... And it grabs the headers for more and more packages as it goes, as each gets referenced from another. When it looks like it is doing nothing, if you watch your internet connection or the spool folder where it keeps the headers, you can see it is still chugging along, feeding itself more and more info. So when you kill it and go to just the first few, it already had most of that stuff from before you killed it, and it doesn't need to go for lots more because you only asked it to do a little this time. I think it would probably eventually finish on it's own if you leave it to sit working on the 130 packages in one shot, but you'd need to watch it to see if it is really still alive. I don't have the patience to wait it out like that! So I end up killing it and going for fewer, like you are doing.