I want to know the maximum number of partitions in a hard drive.

But I'm not sure whether that is affected by factors such as hard drive types or partition scheme or not. So, I'll just write down all possible cases that I can think of.


  1. Does the number of maximum partitions per hard drive affected by
    • partition scheme (GPT vs MBR)
    • The type of hard drive (SCSI/SATA vs IDE - because SATA and SCSI are both treated as SCSI)
  2. What is the maximum number partitions for a hard drive for each case? I think the possible cases include
    • GPT:
      1. max num of partition per SCSI/SATA hard drive?
      2. max num of partition per IDE hard drive?
    • MBR: for MBR, there's the concept of primary, logical and extended partition. logical partitions only exist if an extended partition exists (and there can only be at most ONE extended partition per hard drive). So the questions are:
      1. Does the number of primary partitions affect the number of possible logical partition? (i.e. Does it make any differences to the maximum number of logical partitions if I have 0 or 1 or 2 or 3 primary partitions?)
      2. max num of logical partition per SCSI/SATA hard drive?
      3. max num of logical partition per IDE hard drive?
  • 1
    The actual physical/electrical interface, i.e. IDE, SCSI, SATA, etc. has no bearing on the disk partitioning scheme, i.e. MBR, GPT, VTOC, etc. – fpmurphy May 4 '19 at 14:27
  • 1
    @sebasth I don't see your linked question as duplicate, since the it is limited to MBR partitions, and the answers barely mention GPT at all. – Johan Myréen May 4 '19 at 14:58

Because on some devices partitions are sometimes not used at all, here are some numbers for partitions and their modern replacements;

  • 4 MBR primary partitions
  • or 255 Linux GPT partitions
  • or 16,777,216 LVM1 volumes vg*lv, LVM2 said unlimited but checking the source would likely reveal a limit.
  • or 9,223,372,036,854,775,807 zpools (ZFS) or subvolumes (btrfs)
| improve this answer | |
  • Are these all practical or only theoretical limits? I don't believe you can have that many ZFS subpools or Btrfs subvolumes in practice. – Johan Myréen May 4 '19 at 15:04
  • 1
    The GPT partition table can contain as many partitions as you can handle (billions); I just successfully created 304 partitions on an external disk. On the other hand, the Linux kernel seems to have a limit of 256 on the number of partitions it can use. I don't know if this limit can be raised. – Johan Myréen May 4 '19 at 15:11
  • This answer talks about extended partitions in the EFI partitioning scheme, which that scheme does not in fact have. And the numbers have apparently been pulled out of a hat. – JdeBP May 4 '19 at 16:41
  • @JdeBP I added some links so you need not worry about hats. – user1133275 May 4 '19 at 16:56
  • @JohanMyréen I expect O(1) performance with the 64 bit pointer array. One will encounter other limitations like 64MB is the minimum size, but the point is that there are no limits that are practically reachable in the foreseeable future. – user1133275 May 4 '19 at 16:56

Not the answer you're looking for? Browse other questions tagged or ask your own question.