For IBM PC-DOS/Microsoft MS-DOS, which version was the first version to actually allow using more than one partition for data storage? How about using multiple primary partitions vs one primary plus logical drives within an extended partition? How about multiple drives?

Wikipedia, in its infinite wisdom (err) states that:

The partition table supported up to four primary partitions, of which DOS could only use one. This did not change when FAT16 was introduced as a new file system with DOS 3.0.

The same article states that support for extended partitions and logical drives were added by the time of DOS 3.30, which sounds plausible.

I distinctly recall by the time of MS-DOS 6.x or so having a hard disk that was large enough that I partitioned it into two partitions, so by that time, MS-DOS could certainly use multiple partitions for data and program storage, just not for holding the operating system proper. (Given the size of even MS-DOS 6.x, this wasn't a major concern.) I do not, however, recall the exact way in which that drive was partitioned; it may have been as two primary partitions, or as one primary plus one extended/logical pair.

  • This question was inspired by discussion in the comments to one of the answers to Old school backup systems: what are they?
    – user
    Dec 1, 2017 at 19:09
  • DOS didn't "use" any partition in a way modern OSes do .. and given a core DOS installation is rather small, distributing parts of the OS over multiple partitions would have been extremely nonsensical. Dec 6, 2017 at 16:51
  • @rackandboneman I'm not sure what your point is. Could you clarify what you are aiming at?
    – user
    Dec 6, 2017 at 19:11
  • There is not much of a kernel that uses the filesystem for its own business. There are not many files a DOS system (especially the really old versions) DEPENDS on - io.sys, msdos.sys, a command interpreter, a few optional config files, DONE. Most anything else follows PATH, which is a matter of the command interpreter. If the location of these few files is restricted to the C: partition, this is rather inconsequential. Dec 6, 2017 at 19:42
  • @rackandboneman Sure, but that doesn't negate the value in being able to use additional partitions. Note that the question explicitly states that around the time of MS-DOS 6.x, "MS-DOS could certainly use multiple partitions for data and program storage, just not for holding the operating system proper. (Given the size of even MS-DOS 6.x, this wasn't a major concern.)"
    – user
    Dec 6, 2017 at 20:05

1 Answer 1


The short answer is MS-DOS 3.3.

According to the MS-DOS Encyclopedia, MS-DOS 2.0 to 3.2 (included) only supported one MS-DOS partition per partitioned device (hard drive basically). (Before MS-DOS 2.0, hard drives weren’t supported at all.) MS-DOS 3.3 introduced support for multiple partitions per device, albeit still limited to a single primary partition per drive (and multiple logical partitions inside extended partitions), and famously, 32 MiB per partition.

Compaq DOS 3.31, released in November 1987 or thereabouts, introduced support for FAT16 with 32-bit sector counts and thus partitions larger than 32 MiB. Subsequent versions of DOS included this, starting with IBM DOS 4. (Reportedly, other OEM versions of DOS 3.3 released after Compaq DOS 3.31 included this too, e.g. Zenith DOS 3.3 Plus.) MS-DOS 5 introduced support for multiple primary partitions per drive.

Various manufacturers also provided device drivers which supported proprietary partitioning schemes, or provided support for larger drives using the “standard” DOS partitioning scheme (such as Compaq’s ENHDISK.SYS, mentioned by Raffzahn, which was available as early as Compaq DOS 2.12 if not earlier).

I’m not sure about DR DOS’s level of support; certainly by DR DOS 5 it supported large partitions (I used it on a 120 MiB drive with a single real partition, hosting a compressed volume using SuperStor).

How did MS-DOS assign drive letters in the case of more than two floppy disk drives? covers letter assignment and discusses the differences in more detail.

It might help to put this in context (especially for readers coming from the original comment which led to this question).

  • DOS 3.0 was released in August 1984; the “standard” hard drive then was 10 MB in size, for around $900 (as present in the PC/XT), and the largest drives for sale for PCs that I can find are 33 MB in size (see for example BYTE volume 9 number 9, page 236), which corresponds quite nicely with the maximum 32 MiB usable in a DOS partition.
  • DOS 3.3 was released in April 1987, and $900 then would comfortably buy a 40 MB hard drive; Seagate also had a 80 MB drive, available for around $1300 (large hard drives couldn’t be used in full, simultaneously, with DOS 3.2 and earlier, as standard, but they could be used with other operating systems, and that’s how some were marketed).
  • DOS 4.0 was released in July 1988, by which time Maxtor sold a 130 MB drive for around $1500.

By the time MS-DOS 5.0 was released in 1991, 120 MB drives were starting to become common.

I’ve checked with an old PC and a 40 MB drive, and DOS 3.2’s FDISK can’t create a partition larger than 32 MB, however it can create multiple primary partitions. But only one partition per drive is usable at a time (the active partition). DOS 3.3 can create and use extended partitions with logical partitions inside them.

  • As far as I remember, one didn't upgrade to every new MS-DOS available, usually you only updated it when you really had to. When e.g. your old 20MB HDD was full, and you bought a new shiny 130MB Seagate ST3144 in 1991 or so, you updated from your DOS 3.0 to 3.3 (or 4.0). "Your DOS 3.0 will not be able to use this one, better get a 3.3 or better!" There rarely was any need to update, other than new hardware (at least programs I used didn't require much, like e.g. TurboC ).
    – Tommylee2k
    Dec 7, 2017 at 8:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .