Okay. I have two harddrives. The first with three partitions: WinXP (NTFS), Linux (ext2) and a Linux Swap. My second HD is 300 GB and I want to use it for files, data and everything, and I want it to be visible both for Linux and Windows. I assume I have to use FAT32 for this because Linux only supports NTFS read only. My problems or questions are:

1. I used cfdisk /dev/hdb to make hdb1 a primary partition with partition ID 0C. Is this correct, or should I use 0B? Or even something else?

2. I tried format with "mkdosfs -F 32 /dev/hdb1" and it seemed like a success because I got no error messages (it just returned "mkdosfs 2.10 (22 Sep 2003)") and when I mounted the partition it was empty (I had somehow managed to get a folder there before? o.O ). My HD is 300 GB, but the format took like... two seconds?

3. I assume that the 32 GB limit for formatting FAT32 on WinXP is because as soon as I reach 32 GB used space on the parition there won't be enough clusterspace (?) (I think it was something that is stopped around 32kB) to make use of more dataspace. Is this correct? Is there any way to make this "clusterspace" bigger, like... let's say ten times more (32 GB for 32kb => 320 GB for 320 kB) so I can use my whole 300 GB FAT32 parition?

4. Let's say I have Apache installed on both Windows and Linux. I make them share the same config file that I store on my second harddrive, and I make them share the same www directory (htdocs?). Would this be possible? My goal is so that I can be both logged in on Linux and Windows and still have my server on. I'm pretty sure it is possible, but I'm not sure if it is recommended. I mean are there any bad things doing it?

Thanks.

Your 300Gb baby should work. Fat32 isn't economical in term of disk management.

Fat32 cannot handle a file larger than 4Gb and you will need a NTFS partition if downloading a DVD file.

Got my 200Gb in Fat32 and like you I keep everything in a neutral partition that can be read and written by all systems.

I would probably in your case do 100Gb for NTFS and the rest 200Gb in Fat32. In Linux you can always drag files from NTFS partition into Fat32 partition. If you don't expect to handle large files leave the 300Gb Fat32 alone.

I have drives formatted FAT32 (hex 0c) and there is no issue with sizes over 32 GB.
If you make a new partition and format it, there should not be any folders there--it SHOULD be empty.
Verify what you have with fdisk /dev/hdb, then "p" to list the partitions.

Q3: Don't understand
Q4: Maybe you CAN do this, but it seems that if you want a server running, you would wnat it on all the time---consider putting another computer on and dedicate it as the server. P3 boxes can be very cheap--esp with no monitor.

Thanks. Then I did right with hex 0C I guess. But it only took like two seconds to format it. Could that be possible? I tried format it in Windows yesterday and it took about three hours! (only to be informed AFTER the formatting that it wasn't possible)

I don't need the server on all the time. I mostly only use it to test code (and I don't have a real site to put up there yet anyway).

Hmm.. I'm not sure how I can explain what I meant in Question 2. But as I've understood it there is a space (cluster space I think it is called) where it store pointers to the data in the rest of the partition. When the partition is less than 32 Gig this space is 16 kb I think and when it is more than 32 Gig the space is 32 kb. 32 kb is probably enough for a 32 Gig partition, but mine is 300 Gig and I'm not sure if I will be available to use it all (I don't want it to tell me when I reach 40 Gig "You can't use more of your partition with this filesystem" or something)

There is a quick and slow formats.

In quick format you only install the filing index. The slow way involving zero every byte in the disk. You will find the time taken isn't much different if you "quick format" it in XP. You only need to do a slow format if you use the disk to store the customers accounts in a bank.

I mentioned I used a 200Gb Fat32 disk for my data. Is that is good enough for you?

The test of the pudding is in the eating, so why do you dump a 100Gb of information into the 300Gb disk and see if you can read them, rather than worrying the sky may fall down?

Okay I read more on the cluster thingy and I think the only problem was that every file takes up at least 32 kB (so if I save a text file with just on character (byte) it will still use 32 kB on the hard drive). I can live with that.

Could you explain more about the slow format? Why only for that purpose? Is it because people can still access the information if it hasn't been nulled properly? Is there not anything else bad by using a quick format? I rather make a slow format if it means better things afterwards comparing to a quick format.

In practical term it doesn't matter because according to your filing index the nulled spaces may be null to start in a slow format but you can't "null" every file you delete and so very soon your hard disk will end up with empty spaces not "nulled", or similar to a quick format condition.

The quickest way to use a hard disk is to overwrite whatever information is there. The filing index tells the PC which area is empty even there may be junk information after a deletion.

You only need to "null" the space if you don't want others to know what information was there before but real pros can apparently find it even after you have nulled it several times.

Okay. Thanks for your help. It seems good now so I'll use it.