-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Michael P Gabriel wrote:
> I have been into computers since 1983 when I bought my first TRS80 and
> I have a very high IQ. Yet, I cannot understand FDISK for the life of
> me. I have always used Western Digital's, "LIFEGUARD", for new disks,
> and FORMAT C: with Partition Magic for existing disks.
>
> All I really need is a simple definition of: PRIMARY, SECONDARY,
> LOGICAL and EXTENDED.
> Thanks,
> MIke
> Mike
Note: The following may be easier to read with a fixed-pitch typeface.
'FDISK', according to the IBM PC-DOS 7.0 manual, is short for 'Fixed Disk
Setup Program'.
The 'secondary partition' appears to apply to systems such as the IBM
i270, i820, i830, i840, and i890 systems, so I'll skip that
Hard drives can contain up to four 'primary' partitions per drive.
Under DOS-based operating environments only the 'active', or 'boot'
partition is accessible to the system.
Under 32-bit Windows systems (such as NT/2000 and probably XP), and as I
recall OS/2, a limited number of partition types are recognized. The
partition type is identified by a special code in the 'partition table'
which lives in the first sector (sector zero) of the hard disk.
Systems such as Linux can recognize a large number of partition types. If
support for the recognized types is available in the kernel or as a module,
the system can access the corresponding 'filesystem'.
The partition table also defines the type of filesystem that is expected
within each partition.
Under 32-bit Windows the options are for FAT-16, FAT-32, and NTFS. DOS-based
systems (MS-DOS, Win3.x through Win95 -prior to OSR2) were limited to
FAT-16 and FAT-12.
FAT-12 is still used on floppy diskettes but, as I recall, is no longer
supported under Windows for hard disks.
The size of a partition, and the choice of a filesystem (FAT, FAT-16, NTFS,
etc.) will also determine how much physical space is used to store files on
the hard disk.
// The table below was excerpted from the Microsoft Knowledge Base.
The following is a table of logical drive sizes, FAT (File Allocation Table)
types, and cluster sizes:
Drive Size FAT Type Sectors Cluster
(logical volume) Per Cluster Size
---------------- -------- ----------- -------
(Floppy Disks) 360K 12-bit 2 1K
720K 12-bit 2 1K
1.2 MB 12-bit 1 512 bytes
1.44 MB 12-bit 1 512 bytes
2.88 MB 12-bit 2 1K
(Hard Disks) 0 MB - 15 MB 12-bit 8 4K
16 MB - 127 MB 16-bit 4 2K
128 MB - 255 MB 16-bit 8 4K
256 MB - 511 MB 16-bit 16 8K
512 MB - 1023 MB 16-bit 32 16K
1024 MB - 2047 MB 16-bit 64 32K
// end excerpt
On drives up to 8 gigabytes, FAT-32 uses a cluster size of 4k (4,096 bytes).
Basically this shows that a larger cluster size wastes disk space:
If you have a 2 gig drive (2047 MB above) with its 32K cluster size, any
single file smaller than 32K will eat the entire 32K space.
A 33K file will eat 64K of disk space.
On an 8 gig drive under FAT-32 the same 33K file would only use 36K of disk
space.
NTFS on a 60 gig drive (under Windows 2000 SP3) uses a 4K cluster; on a 15
gig drive it shows a 512 byte cluster size.
I seem to recall that FAT-32 is 32K on large drives although I couldn't
verify that.
For all the DOS-based systems (including ME, as I recall), the system must
either be physically installed on the first primary partition or it must be
fooled into 'thinking' it's on the first primary partition.
Windows NT/2000, XP, OS/2, and Linux can generally be installed on, and
booted from, almost any partition.
An extended partition is useful if more than four partitions are desired on
a single drive.
In a DOS extended partition, you can typically have up to 23 'logical
drives'.
The extended partition essentially becomes a container wrapped around a
collection of smaller partitions.
I believe I got this all correct. Any extra comments are welcome.
Ron n1zhi
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
iD8DBQFAHOWma9fyRcf4bIYRAnD9AKCbqBAJWbvelv3ek1rJmz Lb3NV3WwCeJcPT
fQuaUd1P281TW7Uo93V2CC4=
=IQ41
-----END PGP SIGNATURE-----