Jump to content

Talk:Disk partitioning: Difference between revisions

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Content deleted Content added
Where did the term "partition" originate?: What, did AT&T go with "partition", and then "slice". and then "partition" again?
Where did the term "partition" originate?: Replace your "fff" placeholters before hitting "Publish changes", d00d.
Line 414: Line 414:
::[http://bitsavers.org/pdf/att/unix/System_V_Release_4/0-13-947086-7_Unix_System_V_Rel4_System_Administrators_Guide_1990.pdf The SVR4 System Administrator's Guide], from 1990, uses the term "partition" on page 5-23. On page 5-25, however, they show device names for partitions of the form `cldOsO`; as I remember, the "s" there stands for "slice", which I think was originally a System V-ism. [http://bitsavers.org/pdf/att/unix/System_V_386_Release_3.2/UNIX_System_V_386_Release_3.2_System_Administrators_Reference_Manual_1989.pdf The System V/386 Release 3.2 System Administator's Reference Manual], for SVR3.2 on x86 machines - probably mainly PC-compatible machines - uses "partition", but that may have come from the use of that term on DOS/Windows; I haven't yet found any SVR3 or SVR2 administrator documentation for other platforms, to see whether they use "slice". [[User:Guy Harris|Guy Harris]] ([[User talk:Guy Harris|talk]]) 20:01, 27 January 2022 (UTC)
::[http://bitsavers.org/pdf/att/unix/System_V_Release_4/0-13-947086-7_Unix_System_V_Rel4_System_Administrators_Guide_1990.pdf The SVR4 System Administrator's Guide], from 1990, uses the term "partition" on page 5-23. On page 5-25, however, they show device names for partitions of the form `cldOsO`; as I remember, the "s" there stands for "slice", which I think was originally a System V-ism. [http://bitsavers.org/pdf/att/unix/System_V_386_Release_3.2/UNIX_System_V_386_Release_3.2_System_Administrators_Reference_Manual_1989.pdf The System V/386 Release 3.2 System Administator's Reference Manual], for SVR3.2 on x86 machines - probably mainly PC-compatible machines - uses "partition", but that may have come from the use of that term on DOS/Windows; I haven't yet found any SVR3 or SVR2 administrator documentation for other platforms, to see whether they use "slice". [[User:Guy Harris|Guy Harris]] ([[User talk:Guy Harris|talk]]) 20:01, 27 January 2022 (UTC)
:::Maybe I was thinking of Solaris. --[[User:Chatul|Shmuel (Seymour J.) Metz Username:Chatul]] ([[User talk:Chatul|talk]]) 01:09, 28 January 2022 (UTC)
:::Maybe I was thinking of Solaris. --[[User:Chatul|Shmuel (Seymour J.) Metz Username:Chatul]] ([[User talk:Chatul|talk]]) 01:09, 28 January 2022 (UTC)
::::Well, this is amusing. [http://bitsavers.org/pdf/sun/solaris/2.2/801-4048-10A_Solaris_2.2_Basic_Installation_Guide_199305.pdf FFF] says, in the glossary entry for "slice":
::::Well, this is amusing. [http://bitsavers.org/pdf/sun/solaris/2.2/801-4048-10A_Solaris_2.2_Basic_Installation_Guide_199305.pdf The Solaris 2.2 Basic Installation Guide] says, in the glossary entry for "slice":
::::{{quote|A discrete portion of a disk, configured during installation. Under SunOS 4.l.x and System V Release 3, slices were referred to as partitions.}}
::::{{quote|A discrete portion of a disk, configured during installation. Under SunOS 4.l.x and System V Release 3, slices were referred to as partitions.}}
::::which seems to say that "slice" is an SVR4ism. But, as noted above, the SVR4 System Administrator's Guide use "partition", not "slice". The Solaris 2.2 document says that the device names are of the form {{mono|c''n''t''n''d''n''s''n''}}, where "c" stands for "controller", "t" stands for "target", "d" stands for "device, and "s" stands for... slice, so I think I correctly remembered that as being the convention.
::::which seems to say that "slice" is an SVR4ism. But, as noted above, the SVR4 System Administrator's Guide use "partition", not "slice". The Solaris 2.2 document says that the device names are of the form {{mono|c''n''t''n''d''n''s''n''}}, where "c" stands for "controller", "t" stands for "target", "d" stands for "device, and "s" stands for... slice, so I think I correctly remembered that as being the convention.

Revision as of 03:07, 28 January 2022

Template:Vital article

WikiProject iconComputing C‑class Mid‑importance
WikiProject iconThis article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
CThis article has been rated as C-class on Wikipedia's content assessment scale.
MidThis article has been rated as Mid-importance on the project's importance scale.

Merger

Hi, I returned the materials moved to Hard disk drive partitioning history, because they don't make up for a whole article. Let me explain: each article in Wikipedia has to reach a balance of size versus integration of details. Now the history of hard drive partitioning is quite integral to the concept of drive partitioning, and it doesn't excessively increase the size. So in this case it's best to leave it in one chunk. Remember, Wikipedia is an encyclopedia with proper articles, not Everything2 with nodes :-) --Uriyan

In fact, the history of partitioning is just about the only thing on the topic. The whole concept of partitioning is completely obsolete. Partitioning (and formatting) make no sense in the case of a Logging FS where you can grow and shrink the log dynamically. -- Ark

POV, surely -- partitioning is still critically relevant to the vast majority of filesystems. Wikipedia doesn't enjoy the luxury of closing its eyes and pretending that it doesn't exist. --Jkew 19:46, 22 November 2005 (UTC)[reply]

To do: Drive partitioning on non- IBM PC architectures was moved from the article to here. --cprompt 02:23, 21 Dec 2003 (UTC)

I merged the following articles Active partition, Partition (Computing), Partition (IBM PC) - Jacob 16 Jun 2005

Disk partition and partition table not the same thing. 奧田95 (talk) 11:47, 10 November 2020 (UTC)[reply]

Extended partitions?

Extended partition is linked here, but the article doesn't define - or even mention! - extended partitions, let alone explain how they function. Is this missing for a reason, or does it just await writing? --Dyfrgi 22:42, 7 May 2006 (UTC)[reply]

I added a blurb about this when I rewrote the article a bit. Fiskars007 21:19, 10 August 2006 (UTC)[reply]
What about file systems that don't use boot records? Is the concept of an extended boot record something that concerns PC users, and nobody else? Dexter Nextnumber (talk) 06:42, 4 January 2010 (UTC)[reply]
Yes, the EBR is specific to the PC BIOS partitioning scheme - as is the MBR. Letdorf (talk) 17:44, 4 January 2010 (UTC).[reply]

1024 sector limitation

The article says:

"Technical limitations of a filesystem or operating system (e.g. old versions of the Microsoft FAT filesystem or old Linux kernels that can't boot on a partition with more than 1024 sectors)"

Is this a reference to the old limitation in LILO that meant it could not boot from a partition that began after the 1024th cylinder on a drive? If so, the statement above needs to be corrected. — Yama 10:00, 15 September 2005 (UTC)[reply]

I have removed the dead link to xosl.org (which now belongs to a link farm):

Although Ranish has a mirror of it, I'm not really sure how relevant an operating system loader is to an article on partitioning -- all the other links are to either further information on partitioning or to partitioning utilities, not to loaders. --Jkew 19:30, 22 November 2005 (UTC)[reply]

Extended/logical partitions

24 logical partitions in an extended partition? Since when? Assume this is a typo for 4. Should also note that extended partitions can be nested. --Jkew 19:40, 22 November 2005 (UTC)[reply]

I think this refers to the limit existing since DOS 3.3 (1987) to have a maximum of 23 logical partitions inside an extended partition (using a chained setup where each partition is preceded by an "EBR" with only two entries, one being the logical, and the other being another extended which points to the next slice.)
The limit of 23 was computed as being 26 (letters) -2 (A: and B: were for floppies) -1 (for C: which had to be primary). --AntoineL 15:16Z, December 5th, 2005

  • As AntoineL pointed out, there can be 26 drive letters under DOS/Windows on a PC, so the only typo error was in showing 24 instead of 23, but another OS may be able to create an unlimited number of extended partitions! Linux, e.g., doesn't use drive letters as an IBM/Microsoft OS does.
  • Your comment, "Should also note that extended partitions can be nested." is incorrect, at least according to every OS and utility that I've tested. None of them ever created a nested extended partition! See the discussion page for extended boot record. If you know of a utility/OS that truly creates nested partitions, please list all the details there. Daniel B. Sedory 18:07, 4 August 2007 (UTC)[reply]

Would be useful to provide a link to this for those (like me) who had their interest piqued -- can't find anything suitable within Wikipedia. --Jkew 19:49, 22 November 2005 (UTC)[reply]

Google. ;^ ] 68.64.175.222 22:45, 26 January 2006 (UTC)[reply]

Other partitioning implementations

Concurrent schemes used

There are a number of them; GNU Parted or Linux fdisk support a few like the Unix-inherited (BSD, Sun, etc.) disklabels or the Apple Macintosh scheme; there is also the SCO Xenix divvy scheme which may be historically relevant.--AntoineL 15:51Z, December 5th, 2005

Intel GUID Partition Table

A recent edit (Nov 30th) added

Its however unlikely this project will ever replace IBM PC partitions as the Itanium processor was not received well by the market.

While factually correct about Itanic, I do not agree with the conclusion. For the PC architecture, GPT is still the only scheme with some acceptance which saves the 2TB barrier (32-bit count of 512-byte sectors). It is implemented in Linux, BSD, and the recent versions of Windows.
I am not to say GPT will be the solution (would be a POV.) Booting is certainly still a problem to solve, and the BIOS makers are expected here; also AMD does not seem willingful to acceptance the EFI standard as a whole (but GPT can easilly be severed.)--AntoineL 15:51Z, December 5th, 2005

Too technical

In reading this article, I have come across quite a few terms that would definitely not be readily identifiable to many people who use PCs. Jargon terms are used rather than simple English- defining/wikifying the terms would be a good idea. On the whole, the article is well written from a basic techie viewpoint, but that moves it away from being generally useful to a broad base of Wikipedia readers.

P.MacUidhir (t) (c) 05:34, 26 February 2006 (UTC)[reply]

I agree with you - i actually did try a bit at fixing that up a bit. I find it less confusing now, but it's also noticeably shorter (mostly because the List of partition utilities was given it's own article). Fiskars007 21:19, 10 August 2006 (UTC)[reply]


I was about to try and fix a bit as well, but the oversimplistic Linux-centric view is very unhelpful, and AFAICT, inaccurate, especially all the stuff about how things are partitioned, it would be nice to see Slice vs Partition explanations and where it starts talking about LVM's perhaps a good example as provided by HP (not the Linux copies) would be more helpful than just this stuff about how Linux does it this way or that way. Most people don't use Linux. -- Steve Roome --82.46.105.47 23:53, 1 May 2007 (UTC)[reply]

We already have a slice (disk) article, and BSD disklabel; would a merge make sense? Note that partitions are used by DOS and NT-based systems as well as Linux. –EdC 22:05, 2 May 2007 (UTC)[reply]
We no longer have a slice (disk) article; we have a slice (disk) redirect to this article, with the article now mentioning that "slice" is a term used for disk partitions in Solaris and FreeBSD. Should BSD disklabel be merged here as well? Or should it be left separate, as it's a particular mechanism used to partition disks, just like the MBR mechanism? Guy Harris (talk) 20:27, 22 March 2013 (UTC)[reply]

windows partition setup??

talking about windows partitions... should the programs and files (eg my docs) be stored together on a 2nd partition, or both separately on a 2nd and 3rd?? - mar 1st —The preceding unsigned comment was added by 137.205.148.10 (talkcontribs) 18:20, 1 March 2006 (UTC)[reply]

You should put this sort of question on a computer tech support forum or similar where it may be better answered - this is an encyclopedia, not an IT help desk. If you'd like my recommendation, you would be able to get by with just one or two partitions - adding more partitions makes your life harder, and you run the risk of running out of space for example documents while you still have room for programs (like what often happens to users on Unix-like operating systems). Fiskars007 21:19, 10 August 2006 (UTC)[reply]

One technical detail

The drive mentions making the C drive "nonexistent" and putting the OS on another drive. Could this somehow be clarified? —The preceding unsigned comment was added by 129.7.154.58 (talkcontribs) 00:04, 24 March 2006 (UTC)[reply]

No idea how that works. My computer actually boots Windows XP from H:, not C:, but I still have a C: drive. I think that it's a requirement in Windows still that you must have a C: partition, a holdover from MS-DOS... but I really have no clue. Research should be done on this, or it should be removed. 72.78.224.21 20:48, 10 August 2006 (UTC)[reply]
Often, to create extra space, a hard disk is compressed. When this happens, the compressed data is basically stored as one huge file on the hard disk. It is then opened and treated as a separate disk. To avoid confusion, the drive letters are often swapped, so that the compressed data is C: and the original hard disk (containing the compressed file and usually all or part of the system) is given another letter.130.101.100.106 16:06, 28 October 2006 (UTC)[reply]
This is no longer used by Microsoft windows. DG12 (talk) —Preceding undated comment added 00:50, 11 August 2011 (UTC).[reply]

Needs more info.

On Primary, Logical and Extended Partitions. BKmetic 14:29, 17 May 2006 (UTC)[reply]

I added a mention of them - they should be expanded. I will mark it as a stub. Fiskars007 21:19, 10 August 2006 (UTC)[reply]

UNIX partitions??

According to the article /boot is its own partition. That makes no sense, /boot includes files that is needed to define how to mount the partitions defined in /etc. For these reasons, /boot, /etc, /proc and /bin cannot be their own partitions or the system won't boot. Better be correct. And also, note that very few distributions actually make that making partitions by default. --[Svippong - Talk] 20:49, 23 November 2006 (UTC)[reply]

What you said here is incorrect: I've been booting up my Linux systems for years with a SEPARATE /boot PARTITION. My /boot partition, as do all of them, contains a copy of the Linux kernel, which knows full well how to access whatever data it needs before booting up the whole system completely. You are correct that many recent distributions default to just dumping everything in a single '/' partition, but even under some Live CDs, with some planning, it's possible to install them to more than one partition. Daniel B. Sedory 10:28, 12 April 2007 (UTC)[reply]

Partitioning = Dissection? Not always...

"In layman's terms, partitioning a hard drive makes it appear to be more than one hard drive, especially in how each partition is formatted for different operating systems, and in how files are copied from one partition to another."

Not true. Partitioning is not JUST subdividing a physical hard drive into multiple logical partitions, even creating one partition on one hard drive is still partitioning, just not as advanced. Also, partitioning several disks (such as disks in a RAID format) still do not involve taking one hard drive and dissecting it.

Basically, since I can't think of how to reword it, someone else think of it. (This is in the intro)

--DEMONIIIK 01:13, 9 April 2007 (UTC)[reply]

How about one of these:
Partitioning a hard disk reserves some specific area within it so that an operating system may use that space as a whole logical drive or as part of a larger file system. A single hard disk may thus contain one or more logical drives or even just part of a logical drive that could span several hard disks.
or
Partitioning a hard disk defines specific areas (partitions) within the disk so that they may be used by one or more Operating Systems. Any one partition may form part of a larger file system or it may be an entire logical drive.
or
Partitioning a hard disk drive defines specific areas (the partitions) within the disk. A partition may constitute an entire logical drive or it may form part of a larger logical drive which could span over several partitions and hard disks.
Intersofia 15:31, 22 May 2007 (UTC)[reply]

Active partitions and list of file systems

Could anyone give an explanation —The preceding unsigned comment was added by 193.112.172.12 (talkcontribs) 15:16, 28 February 2007 (UTC)[reply]

It appears that trying to squeeze gigs of data into a CDROM will give rise to an active partition, up until the CDROM becomes full, after which the partition probably becomes inactive.
As for a list of file systems, that would be outside the scope of this article's main page. Best to click your mouse on the link for file systems, and ask over there. Dexter Nextnumber (talk) 01:16, 5 January 2010 (UTC)[reply]

Partitions are not limited to hard disks

Correct me if I am wrong, which is why I haven't waded in to correct the article. A partition is a way of dividing up space in a volume. A Volume can be anything from a file on a filesystem already present to a real hard disks volume, in some situations a volume can span a number of hard disks or other storage media. So as I see it the hierarchy is a follows:

Physical Storage media
>Volume
>>Partition Table (or system specific equivalent)
>>Partition(s)
>>>>Filesystem
>>>>>>File

It's more complicated than that (of course!). There are various ways of slicing and dicing disks, and partitioning is only one of them. You basically do one of of two things:
  • Join volumes together (either to make more space, or more redundancy):
    • Mirroring (RAID 1)
    • Striping (RAID 0)
    • RAID 5
    • Disk concatenation
    • Logical volume groups
  • Divide a volume into pieces (for ease of administration)
    • Disk partitioning
    • Logical volumes
So you might do something like this with 4 100GB disks:
  1. Partition disks up into 4*50GB and 2*100GB
  2. Mirror equal sized partitions to get 2*50GB mirrors and 1*100GB mirror
  3. Make a volume group out of 1*50GB mirror and 1*100GB mirror to give you a 150GB volume group
  4. Divide the volume group into 2*75GB logical volumes
  5. Use the 50GB mirror and 2*75GB logical volumes as you want
I haven't even mentioned hardware RAID, SANs etc. The disks mentioned in step 1 could actually be made of multiple disks striped together in hardware (Do NOT do this! Mirror first, then stripe), or be network SAN volumes that have themselves been sliced and diced from a pool of hundreds of disks. Storage management is getting more and more abstract, with more and more layers. What is a "physical disk" in one layer is just a "logical volume" in the layer below.192.171.3.126 (talk) 17:15, 30 March 2009 (UTC)[reply]

So the common usage of partitioning the hard disk is not strictly true.

And the term Volume comes directly from the literal meaning for space.

Interestingly on some systems, (GNU/Linux) included, one can write a filesystem directly to a block device and skip the partitioning altogether, a filesystem directly on a volume. In this case one does not, for example, mount "/dev/hda1" but the equivalent would be "/dev/hda".

With all of the references to hard-disks in this article it is a big edit and one that I would appreciate some feedback on beforehand.

Thanks. 81.138.5.131 (talk) 14:29, 11 February 2008 (UTC)[reply]

User 81.138.5.131, actually, it depends upon the system used to do the partitioning! Under any DOS OS or even many UNIX/Linux 'fdisk' programs, or anything that Microsoft now refers to as "Basic Disks," some of your comments above would definitely be in error.
But before I go into that, let me say I'm glad you brought this up, since most servers no longer use simple "Basic Disk" configurations, and there should be an article which addresses what is commonly found on such systems! However, the present article has always (so far at least) reflected "Disk partitioning" on Basic Disks only. We need to ask some very experienced AND computer savvy Wikipedia editors if they believe the topic should be covered under different articles (maybe such articles already exist!), or if the present article should be altered to include both "Basic" and what Microsoft calls "Dynamic Disks."
As to your comment: "A partition is a way of dividing up space in a volume." That is incorrect! The term "volume" or "logical volume" (at least under "Basic Disks"), is either equivalent in size to a partition or smaller than a partition, but you never divide a volume into partitions. A primary NTFS partition (on a "Basic Disk") will always be one sector larger than the Volume it contains. It is not the Volume boot record sector, but rather a 'backup' boot sector that it stores in the last sector of the partition. And all extended partitions, of which there should be only one per disk, will contain one or more "logical drive volumes" within it.
When we move to "Dynamic Disk" schemes, a logical volume can 'span' across more than one physical hard disk! Will discuss this further at a later date. Daniel B. Sedory (talk) 22:11, 11 February 2008 (UTC)[reply]

I have just a few minutes to add a bit more: You can define any word however you want, but that won't change how it's already being used by computer industry leaders, techs, journals, etc. This made me search Wikipedia for its use, and I found the article: Volume_(computing). It does point out that volumes can exist without partitions and that volumes can span physical disks, etc., but begins by describing a volume as: "a single accessible storage area with a single file system, typically (though not necessarily) resident on a single partition of a hard disk."

As to your comment, "one can write a 'filesystem' directly to a block device and skip the partitioning altogether," is this really any different than writing a raw image file to a floppy or hard disk drive under a Windows OS?! Unless you meant something entirely different, the 'file system' you speak of would still have to be formatted or even partitioned first and then formatted before being written to the disk. Perhaps you meant how some devices, such as a USB stick, etc. can easily be set up as a single volume (no MBR sector)? But these too could also be partitioned if I wanted them to be (except for using a diskette in a real floppy drive). The main point in all of this is that partitions are never within a volume, but the opposite is quite often true! Daniel B. Sedory (talk) 10:54, 13 February 2008 (UTC)[reply]

To me, «write a filesystem directly to a block device» is unmistakable — he literally means formatting the block device. I am using a harddisk formatted this way right now!
Example: Let /dev/sdx be a harddisk (of the type BIOSes expect to find a partition table on).
Then, noone stops you from doing `mkfs.ext4 /dev/sdx && mount -t ext4 /dev/sdx /mnt/MyEntireHarddisk`.
Pros/Cons: You won't be able to boot from a harddisk formatted this way, and partitioning tools will mistakenly assume that unformatted==unused space, but the zero filesystem offset satisfies every thinkable alignment (ideal for SSDs).
Let's define "volume" as where you can put a filesystem:
  • On Windows, that would be only the inside of partitions, and Mr Sedory would be correct.
  • On UNIX, storage media and partitions are interchangeable as block devices. Citing Logical Volume Manager (Linux): "The term "volume" refers to a disk drive or partition thereof."
Anordal (talk) 00:15, 13 February 2013 (UTC)[reply]

"Disk partition" versus "disk partitioning"

WP:VERB says "do not use verbs for article titles if there is a more appropriate noun title". It's true that, when initially creating partitions, one describes it as "partitioning the drive". However, discussion and documentation seems to use the noun form much more frequently... primary partition, partition table, swap partition, etc. In my opinion, the article should be renamed to "disk partition". --Underpants (talk) 04:27, 14 June 2008 (UTC)[reply]

Both words, "partition" and "partitioning", can be nouns or verbs, depending on their use in a sentence. Since an article title isn't a sentence, there's no way to determine which part of speech is intended. Giving Wikipedia the benefit of the doubt, I'd consider them both nouns when used as titles of articles, so WP:VERB doesn't apply. Unfree (talk) 17:31, 13 August 2009 (UTC)[reply]

Halve

From the article: "This partition is expected to: occupy the outer tracks of the hard drive, more than double the throughput, more than half the access time."

I suspect the editor intended to say "more than halve". I'd change that, considering that more than half isn't as good as less than half, which was probably intended, but what about the verb-noun confusion? Unfree (talk) 17:20, 13 August 2009 (UTC)[reply]

Block size

Interesting but not included. Yet. --88.148.208.54 (talk) 12:22, 30 August 2009 (UTC)[reply]

Overflows and error recovery?

I tried to read the main page to this article and remain confused.

There are lots of file systems that support the concept of a "partition" but the main article seems to proceed on the theory that the user is using (and is familiar with) PC compatible harddrives, on which many partitions may be found. But I am not a PC user, and don't understand how PC harddrives work. I have only theory to fall back on, and since I don't have a PC compatible, I find their method of solving the overflow problem to be less than interesting. In fact, most of the article on that account could be deleted because it is not particularly notable to those who do not own PCs.

How does a typical file system usually deal with an overflow condition? Does the partition automatically expand, as by automatically editing itself, and thus getting bigger? Or does the file system automatically create a new partition, and deposit the overflow there? Dexter Nextnumber (talk) 06:33, 4 January 2010 (UTC)[reply]

Typically, nothing happens when a filesystem becomes full. Resizing and creating new partitions or filesystems is usually a manual intervention. Letdorf (talk) 17:52, 4 January 2010 (UTC).[reply]
So the data is lost? Dexter Nextnumber (talk) 00:39, 5 January 2010 (UTC)[reply]

Depositing error codes?

Is there a way of determining which partitions receive file read/write errors? And if there are, which partition receives the error codes? Or errors relating to misused channels? Is this sorted in any way? Is there at least one file system that automatically creates new partitions when old partitions are full, or when they approach a near-critical density? Dexter Nextnumber (talk) 06:39, 4 January 2010 (UTC)[reply]

Cylinder and volume allocations?

For those of us who are much more familiar with BAMs than FATs, how many file systems define their partitions in terms of allocated cylinders and volumes? I can't be the only one who thinks the FAT system is way too clumsy to be of any practical use. Dexter Nextnumber (talk) 00:49, 5 January 2010 (UTC)[reply]

what is partation

ahmad shoaib ahmad khan say that partation is the deviding the harddisk —Preceding unsigned comment added by 119.73.4.192 (talk) 11:28, 26 February 2010 (UTC)[reply]

Benefits

"Having an area for operating system virtual memory swapping/paging." I'm not saying this is wrong, but is this an actual benefit? Just about every disk, even one with a single partition, already has such an area. If the disk is partitioned, then the page file will be limited to one partition. One theoretical advantage is that the read/write heads will not have as far to travel, as mentioned under "Short stroking". On the other hand, the page file may become fragmented, forfeiting some of those advantages.

Is there any objective evidence showing that it is faster or otherwise better to place a page file on a separate partition? Marzolian (talk) 20:31, 11 May 2010 (UTC)[reply]

What this statement is referring to is a swap partition, ie. a raw partition (without a filesystem on it) used entirely for swap space. Older UNIX and UNIX-like OSs didn't support swap files, so swap partitions were the only option. For systems that support both, in theory, a swap partition should be faster, as pages are read and written directly from/to the disk device, without the overhead of filesystem I/O. In addition, some UNIX systems can use a swap partition for temporary storage of system crash dumps in the event of a kernel panic. See, for instance [1]. Letdorf (talk) 21:41, 11 May 2010 (UTC).[reply]

Rename without "Disk"

Since this applies to general secondary storage (e.g., SSDs, etc.) and not just disks (e.g., HDDs), I propose renaming the article to not contain the word "disk". Perhaps there are better ideas but my initial choice is "Drive partitioning". Uzume (talk) 16:09, 23 May 2010 (UTC)[reply]

Nomenclature is tricky here. While SSDs contain no circular components, and hence are not "disks", they also contain no moving parts, and hence are not, strictly speaking, "drives" either. To be pedantic, this article should probably be called "Secondary data storage device partitioning", but I think that might be going too far! --Letdorf (talk) 12:00, 24 May 2010 (UTC)[reply]
IBM faced this issue in the 1960's, and coined the term DASD as a generic term for referring to data cells, disks and drums. Why not use that here? Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:16, 31 December 2010 (UTC)[reply]
DASD is very IBM-specific and not widely used outside that context, IMHO. Letdorf (talk) 22:09, 3 January 2011 (UTC).[reply]
"Secondary data storage device partitioning" would go too far indeed :-). I would go for "Drive partitioning", although strictly speaking not correct. But you cannot just rename the article, because some details are only about HDD and not about SSD or USB or ... Eg. I don't know if an SSD also works with an MBR? --Mattias.Campe (talk) 14:19, 19 January 2014 (UTC)[reply]
I don't think you partition tape drives, optical drives or floppy disk drives so "Drive partitioning" seems out. I'm not sure u can partition a USB thumb drive either. Since SSDs generally present themselves to the system as an HDD I suggest we leave the title alone and deal with SSDs with a footnote if at all. After all in Windows SSD partitioning is under Disk Management. Tom94022 (talk) 07:00, 20 January 2014 (UTC)[reply]
"I'm not sure u can partition a USB drive either." You can - I just did. (I didn't really need that ancient Quicken backup anyway. :-)) 4 partitions. And we're talking "USB drive" in the sense of "USB thumb drive", not "USB-attached HDD", which I'm certain you can partition - the OS, or, at least, most if not all OSes, don't care. USB-attached drives, like Firewire-attached drives and Thunderbolt-attached drives, present themselves to the system, except maybe at the lowest layer, the same way HDDs do; whether I'd call that "as an HDD" is another matter. Guy Harris (talk) 07:19, 20 January 2014 (UTC)[reply]
Thanks for the clarification. To be even more clear, as I understood it when I wrote the above, USB thumb drives cannot be partitioned under the Microsoft supplied Windows OS since the drive is identified as removable media; however if the removable bit is changed so that the drive is recognized as non-removable media then it can be partitioned like any normal hard disk drive. What I was not certain about was whether there were ways other than this hack to do so, particularly in OSes other than Windows. I'm still not sure :-) But that really doesn't matter since even if USB thumb drives can be partitioned by some means that doesn't mean all drives can be - can a "Firewire-attached [tape] drive" be partitioned? Therefore the proposed title change remains IMO inappropriate. Tom94022 (talk) 17:12, 20 January 2014 (UTC)[reply]
The means I used to partition the thumb drive was Disk Utility, the same app you use to partition any other disk on OS X, so, yes, you are now sure that there are ways other than a hack to partition USB thumb drives - or other any other drive externally attached via USB/Firewire/Thunderbolt/etc. - like other drives on at least one operating system.
The key technical difference isn't whether it's called a disk or not, the key technical difference is whether it's a random-access device or not. If somebody built, for example, a USB DECtape controller offering a random-access-device interface, and plugged it into a box running OS X (or, I suspect, at least some other non-Windows personal computer operating systems), you'd be able to partition it if it had enough blocks to be partitioned and formatted on that OS. Unfortunately, I'm not sure there's a good term to use other than "disk" for that; as noted above, the IBM mainframe term "Direct Access Storage Device" is "not widely used outside that context", so calling the page "DASD partitioning" would probably confuse most readers. If somebody can suggest a term that would be understood by most people and that covers random-access devices regardless of whether they have spinning disks coated with magnetic material or not (and whether or not they're attached via an "external" bus or not), that would be great. (I'm not sure about "drive" either, for no other reason that, for drives with removable media, you're partitioning the medium, not the device into which you insert the medium.) Guy Harris (talk) 19:40, 20 January 2014 (UTC)[reply]
Thanks for the info about OSX. FWIW, I don't think partitioning is necessarily limited to random access; one could partition a tape medium - but I don't know of any OS that does. Since OSX uses "Disk Utility" to partition USB Thumb drives can we now agree that the article title does not need to be changed? Tom94022 (talk) 07:36, 21 January 2014 (UTC)[reply]

Scope of article?

The current text of the article discusses partitioning of disks on a PC, where the partioning information is carried on the disk itself. Is partitioning software where the data structures are external to the disk in scope, e.g., minidisks in CP-67 and VM? Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:16, 31 December 2010 (UTC)[reply]

Yes, I'd say partitioning systems where there is no on-disk partition table should be included, if reliable sources can be cited. Regards, Letdorf (talk) 22:17, 3 January 2011 (UTC).[reply]

lossless resizing?

I'm just wondering that there is virtually no Software known that declares to lossles resize partitions. For now i only see FIPS and ntfsresize. (Both with quite limited scope) --Itu (talk) 00:52, 9 January 2012 (UTC)[reply]

For http://www.lissot.net/partition/partition-03.html message: The domain name you have requested isn't available --92.36.130.23 (talk) 16:30, 22 February 2012 (UTC)[reply]

For http://www.lissot.net/partition/partition-03.html message: The domain name you have requested isn't available --92.36.130.23 (talk) 16:30, 22 February 2012 (UTC)[reply]

Not destroying users files when parititions a disk

This article can be improved. I have C drive. However, I want to split this into two. However, I want to preserve the file data. I think a program like Partition Magic, seem to be able to create a partition without loosing data. However, there are a lot of programs out there, which do not preserve existing data on the disk.

can there be additional information about programs which preserve user files versus programs which do not?

Thx,

Wikipedia is not a manual, guidebook, textbook, or scientific journal. There could be such information, but relying on Wikipedia as the place to get such information is unwise. Guy Harris (talk) 21:38, 8 April 2014 (UTC)[reply]

The definition of partitioning

There seems to be some confusion as to what the definition of partitioning actually is among some editors of the article. If you check the refs here[1][2][3], you'll notice that the explicit definitions for partition in general usage and in computing usage are different. A partition in computing does not necessarily entail a division of space; it is, rather, "a logical section of a disk", or "a section of storage space on a hard disk". Please note that the two are not the same.

A logical container is what a partition is. You can have just one partition on a hard drive, and in fact, you must have at least one partition per hard drive (depending on OS). None of that echoes a division of area. You're just confusing what it can be and what it actually explicitly is.

Hopefully this wasn't too long-winded and got my point across clearly and succinctly.
--

BlueFenixReborn (talk) 05:20, 3 May 2015 (UTC)[reply]

Thanks for attempting to explain but I really don't get your point. None of the cited references support the term "container" but do use terms such as divide or section, either of which is preferable to "container." The references are redundant and two are not particularly relevent to this article. The article is about partitioning, not formating - formating may creat a first partition (I'm pretty sure not all OSes have partition tables) but partitioning implies segmenenting or dividing storage into more than one particition. You have a strong POV but nothing in what you cite supports the use of container in this article. Here is probably the most authorative definition of partition, from the IEEE Dictionary, "A portion of a storage medium that is treated as a if it were an individual medium, as in a partition of a hard disk." No container. Even your definitions support other language than container, i.e.
  • "a logical section of a disk", is created by dividing or segmenting the disk or
  • "a section of storage space on a hard disk" is created by dividing or segmenenting the disk.
in both cases using dividing or segmenting or sectioning with their ordinary english language meaning whether or not these two are the same. Finally, I really don't understand why partitioning isn't a division of area, or more precisely storage volume; you have several times asserted such without any justification other than yr POV. There is simply no justification for introducing an new term not supported by any reference and not particulary revelent to a readers understanding. Tom94022 (talk) 19:39, 4 May 2015 (UTC)[reply]
I understand that from all the references stated here thus far, it does not seem as if the word partition can mean something that is not a division of area; the way they have written it, surely they would have added an explicit note that the "logical section" can equal the entire span of the drive, right? As a result, things are somewhat shaky and ambiguous right now, and it would certainly seem that "partition" would mean a division of area and only a division of area.
I also admit that my knowledge in this area is not as up to par as I would like it to be; I think, actually, that my confusion stems from Microsoft's literature. They seem to refer to single volume VBR-laden devices, without an MBR, as partitions. (link here, under What are partitions and volumes?) So in the end it's an issue of semantics I think. Sorry about editing this, but that actually doesn't make any sense. On Windows at least, you are always able to create/resize/delete partitions, even if you installed it onto a physical media without dividing it into multiple partitions. That suggests Windows always installs an MBR. Further, it means that you can have a single partition on an MBR-laden device. It wouldn't make any sense to start calling the same code (partition demarcations) different things now, would it? I therefore conclude that the definitions cited here, in especiality the IEEE one, do state it with the knowledge that said logical section can equal the entire span of the drive.
Which way do we go on this? -- BlueFenixReborn (talk) 05:48, 5 May 2015 (UTC)[reply]
I agree with your now stricken comment that this is a semantics issue. The article should not be limited to what Microsoft does; it has a few of the many File Systems and has done different things at different times. The article should focus on what "Disk partitioning" is when used in a computing context. My objection to the current edit is the use of a new and unnecessary term "container" that has no support in any reference and adds nothing over the more commonly used terms such as divide or segment.
It is clear that a single storage volume can be one partition, I think the action verb "partitioning" is used to denote the creation, deletion or rearrangement of partitions. Formating is the action verb for creating the first partition in File Systems that support them. The article is about "Disk partitioning" not "Disk partitions" Tom94022 (talk) 08:46, 7 May 2015 (UTC)[reply]
There is no disagreement that one logical section can equal the entire storage medium; I just don't think and so far there are no reliable sources that creating such is "partitioning" other than perhaps the degenerative case where all partitions are eliminated.
Here are some reliable sources that support:
That covers most file systems by installations. Enuf said. I'm going to make another attempt to revise the lede - please do not just revert but try instead to improve. Tom94022 (talk) 08:46, 7 May 2015 (UTC)[reply]
Yeah, I see now. You're right: the problem actually stems from people's inconsistent usage of language. Instead of using the word "partitioning" as the verb form of the specialized computer term "partition", people seem instead to mean "partitioning" in the general sense of the word - namely, a division. Dumb.
So as a side note, the Microsoft ref I tossed in here actually agrees with the paradigm - it uses the noun form, "partition", so everything is consistent. -- BlueFenixReborn (talk) 04:17, 8 May 2015 (UTC)[reply]
  1. ^ Partition, as referenced by The Free Dictionary by Farlex at American Heritage Dictionary of the English Language, Fifth Edition. Accessed April 30th 2015
  2. ^ Partition as referenced by Dictonary.com at The Free On-line Dictionary of Computing, © Denis Howe 2010; under subsection partition in Technology. Accessed April 30th 2015
  3. ^ Partition at FOLDOC, 1996-12-06. Accessed April 30th 2015

Primary partitions limit

The "PC partition types" sections currently states that the "total data storage space of a PC HDD can contain at most four primary partitions, or alternatively three primary partitions and an extended partition." This only applies to MBR (as confirmed by the sentence that follows). Is seems that most new computers now come with GPT instead of MBR. Shouldn't there be a mention that GPT has no such limit and no notion of extended partitions? 24.202.251.80 (talk) 18:48, 12 September 2016 (UTC)[reply]

I added clarifications to avoid misleading readers into thinking that MBR is current technology. 24.202.251.80 (talk) 00:45, 19 September 2016 (UTC)[reply]

Hello fellow Wikipedians,

I have just modified one external link on Disk partitioning. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 22:35, 13 December 2016 (UTC)[reply]

Poorly-written article

This entire article should be scrapped and rewritten by someone who isn't an "armchair computer expert" with no technical knowledge of non-PC systems.

The fundamental error being propagated throughout this article is the equating of a partition with a filesystem. In the Linux and Windows world, that is often the case. However, a partition and a filesystem are fundamentally different structures. The former is an arbitrarily chosen fraction of the total raw disk space. The latter is the organization of some or all of that fraction of total disk space.

Using x86 architecture hardware as an example, the bootstrap code in the machine's firmware has no knowledge whatsoever of partitions or filesystems. It only knows enough to load physical sector zero (the "boot block") of the selected boot device—which is not necessarily a disk drive—into core at some agreed-upon location, examine the core image for a starting vector and then execute the code pointed to by that vector. That code, in turn, could examine the boot sector image for a partition table, search the table for a partition in which second stage boot code would be found, load that code from the boot device and execute it, and so on.

Or, the boot block could contain only a basic bootstrap program that merely loads subsequent blocks from the boot device and treats them as a program, a practice that was quite common with proprietary minicomputers of days gone by. In other words, partitioning is not required to use a disk, build a filesystem or load an operating system. Partitioning is a convenience that allows more than one operating system to be present on the same storage medium.

A filesystem can exist on a disk that is not partitioned, as a filesystem organizes raw space, which in this case would be part or all of the entire disk's raw space. There have been many computer systems that had no partitioning of any kind on their disks. Try thinking outside of the PC world for a change. — Preceding unsigned comment added by 38.69.12.5 (talk) 20:48, 4 April 2017 (UTC)[reply]

Oh, heck, you don't even need a file system. In the old days in the IBM mainframe world, disk storage was simply assigned to site-defined "data sets" (files) in terms of ranges of cylinder numbers, managed by the site. There were no files or directories or anything else you'd recognize as a "file system". Jeh (talk) 20:33, 18 July 2018 (UTC)[reply]
Jeh, the IBM/360 OS and DOS disks had file systems. The metadata were: volume label (~ superblock), VTOC (~ directory) and file label for each file. The "data set" was the official term but anyone called it "file".

78.29.74.199 (talk) 09:45, 29 November 2018 (UTC)[reply]

"Advantages" and "Disadvantages" sections

These things have existed completely unreferenced for over five years and been tagged as OR/CN for almost two. I'm deleting them. I will delete them again if they're restored without a reference to a RS for every claim. Jeh (talk) 20:40, 18 July 2018 (UTC)[reply]

Interesting that you deleted the entire section but left the list in the unix-like systems which actually applies to any partitioning. Before I noticed your edit I felt it necessary to add "why bother" to the introduction paragraph. Perhaps a better approach on you part would be to research various claims and add references. Maybe some of the original discussion. should be restored. Sincerely, DGerman (talk) 00:42, 12 February 2019 (UTC)[reply]

Where is incompatibility?

"The Windows Disk Manager in Windows Vista and Windows 7 utilizes a 1 MB partition alignment scheme which is fundamentally incompatible with Windows 2000, XP, OS/2, DOS as well as many other operating systems"

Any new and old OS may use not-aligned partitions. For devices with physical block size is 512 it works perfectly. For devices with larger physical blocks it still works but with decreased performance. And older OS will also work with new aligned partitioning and even get advantage of better performance.

The only incompatibility is possible if offset 63 sectors for the first partition was hardcoded by someone (too silly).

78.29.74.199 (talk) 08:59, 29 November 2018 (UTC)[reply]

Proposed merge of Partition table into Disk partitioning

The stub Partition table does not contain enough content to merit a second article. Ethanpet113 (talk) 11:18, 20 September 2020 (UTC)[reply]

Agree It's been forever since someone replied to this, so, support, and I'm gonna go ahead and merge it. Feel free to revert me if I'm wrong. WhoAteMyButter (📨📝) 21:18, 26 March 2021 (UTC)[reply]

Definition of partition in this article is wrong

A partition is a partition, it is the divider between spaces, not the actual space. 218.214.190.174 (talk) 10:45, 1 January 2022 (UTC)[reply]

In the computing field, it's the space. In this case, the divider isn't a physical object taking up space (a wall between two rooms takes up space that belongs to neither room, but that's not the case for partitions in primary or secondary storage). See, for example, the Lexico definition for UK English and the Lexico definition for US English which list, as definitions 2.1 and 1.3, respectively, for "partition", "Each of a number of portions into which some operating systems divide memory or storage." as the definition in "Computing". Guy Harris (talk) 18:28, 1 January 2022 (UTC)[reply]
When studying computer science we learned a partition is what you put between two logical drives. If you think it is the actual logic drive, how many partitions do you have by default on your floppy disc? 218.214.190.174 (talk) 3:28, 27 January 2022 (UTC)
I don't have floppy discs any more, but, when I did, they had one by default, taking up the entire drive. Guy Harris (talk) 01:07, 27 January 2022 (UTC)[reply]
My recollection is that under MS and PC DOS FDs could not be partitioned. Regardless, Microsoft's 1988 MS-DOS Encyclopedia (p. 90-91) clearly describes and depicts partitions as the space allocated and not the divider. Tom94022 (talk) 07:38, 27 January 2022 (UTC)[reply]

Where did the term "partition" originate?

AFAIK the earliest use of partition in a computing environment was in 1983 with IBM PC-DOS 2 which first supported HDDs and introduced the FDisk command. It could support up to four "partitions." My recollection is that the term was later adopted by Apple, the various UNIXs and even IBM mainframe OSes. Anybody have any earlier usage? Probably should add something about its origin to this article. Tom94022 (talk) 07:52, 27 January 2022 (UTC)[reply]

The UNIX V6 man page for the RP-11/RP03 device and driver speaks of the disk being divided into "sections", "pseudo-disks", and "portions". The UNIX V7 man page for that device and driver uses "sections", "pieces" and "pseudo-disks". No "partitions" at that point. The rp page in the System III manual, from 1980, uses "sections" and "pieces". The same applies to the RM80, RP03, and RP07 pages in this UNIX 5.0 (a/k/a System V) manual from 1982.
The UNIX V8 man page for the RA60/RA80/RA81 MSCP devices and driver uses "portions", "pseudo-devices" - and "partition". The System Manager's Manual for the Sun Workstation - Models 100U/150U, from September 1983, also refers, in Appendix B, to "partitions".
So, at least for UNIXes, the use of "partition" may have come from DOS. Guy Harris (talk) 10:38, 27 January 2022 (UTC)[reply]
That depends on the context. IBM used the term Partitioned data set for OS/360 libraries and used the term partition for storage allocation in DOS/360 and OS/360 MFT.
In the context of virtual disk drives, PCDOS is the first I'm aware of. I believe that BSD used the term slice. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:04, 27 January 2022 (UTC)[reply]
The 4.2BSD hp man page from 1983 (for MASSBUS disks; nothing to do with Hewlett-Packard) uses "partitions".
The SVR4 System Administrator's Guide, from 1990, uses the term "partition" on page 5-23. On page 5-25, however, they show device names for partitions of the form `cldOsO`; as I remember, the "s" there stands for "slice", which I think was originally a System V-ism. The System V/386 Release 3.2 System Administator's Reference Manual, for SVR3.2 on x86 machines - probably mainly PC-compatible machines - uses "partition", but that may have come from the use of that term on DOS/Windows; I haven't yet found any SVR3 or SVR2 administrator documentation for other platforms, to see whether they use "slice". Guy Harris (talk) 20:01, 27 January 2022 (UTC)[reply]
Maybe I was thinking of Solaris. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 01:09, 28 January 2022 (UTC)[reply]
Well, this is amusing. The Solaris 2.2 Basic Installation Guide says, in the glossary entry for "slice":

A discrete portion of a disk, configured during installation. Under SunOS 4.l.x and System V Release 3, slices were referred to as partitions.

which seems to say that "slice" is an SVR4ism. But, as noted above, the SVR4 System Administrator's Guide use "partition", not "slice". The Solaris 2.2 document says that the device names are of the form cntndnsn, where "c" stands for "controller", "t" stands for "target", "d" stands for "device, and "s" stands for... slice, so I think I correctly remembered that as being the convention.
So I have no clue what the history of the terminology for disk pieces is in System V - when did they use "partition", when did they use "slice", and when did they use something else? Guy Harris (talk) 03:05, 28 January 2022 (UTC)[reply]
CP/VM use “minidisk” for the same concept. Peter Flass (talk) 02:47, 28 January 2022 (UTC)[reply]