Talk:NTFS/Archive 2

ntfsprogs

ntfsmount is a part of ntfsprogs. Why don't we replace "ntfsmount" with "ntfsprogs"?. There is also other programs in the ntfsprogs project. For instance, there is ntfsresize. --Ysangkok 20:39, 20 April 2007 (UTC)

What is Difference between NTFS & FAT32..

What is Difference between NTFS & FAT32.. does it effects the speed for those who use heavy video conversions —The preceding unsigned comment was added by 125.99.255.26 (talk) 12:00, 12 May 2007 (UTC).

FAT32 is faster for the same cluster size unless you have too many files. Bigger cluster size is better. But for very large disks, even if you only have one file, you have to use NTFS, so you may not have a choice. NTFS has more features. NTFS handles folders with thousands of files faster. Big hard disks are always faster than small hard disks, so a very large hard disk with very large clusters will be faster than your old disk even if your old disk has FAT32.—Preceding unsigned comment added by 218.214.18.240 (talk • contribs) 11:49, June 10, 2007

ALSO fat32 cant handle single files over a certain size (4gb?) whereas NTFS can handle much bigger files. so if you are handling huge video files , NTFS is better. —Preceding unsigned comment added by 86.16.160.17 (talk) 10:29, 9 June 2008 (UTC)

Reliability of Paragon's NTFS for Linux driver

There is/was a sentence which reads: "NTFS for Linux: Full write support is available using Paragon's NTFS for Linux driver, although criticised for leaving many errors on the volume when mounted read-write.[citation needed]" And naturally enough someone requested a citation. I've looked around and could not quickly (5 min) find any supporting evidence for criticising the driver for leaving many errors. On the contrary I found an article ['In the Window' (166KB PDF] from http://www.linux-magazine.com dated January 2007 in which they praise the driver and it's documentation and utilities, though questioning the price. So I am going to remove the critical part of that paragraph. If this decision is wrong please provide at least one well-informed citation before reverting. --Duncan nz 10:29, 5 June 2007 (UTC)

NTFS has 5 versions

In what sense did it have 5 versions, if 1 and 1.1 were never released, and are not discussed? Or why not 6 versions, wasn't there an equally undistinguished version 2? —Preceding unsigned comment added by 218.214.18.240 (talk • contribs) 11:36, June 10, 2007

macFUSE

can someone explain how macFUSE would help me write from my macbook pro (intel chip) to my NTFS external? I saw that as an option somewhere but I don't see "macFUSE" listed any where in the article after doing cmd+f for "macfuse." Thanks. Tkjazzer 03:45, 15 June 2007 (UTC)

Replaced FAT?

Should that statement be ore specific? I'm not sure if NTFS replaced FAT16 entirely or the FAT format. NTFS is only used in Windows NT. I will look into that. A Raider Like Indiana 19:34, 17 August 2007 (UTC)

NTFS replaced FAT as the default filesystem in Windows. But not in general. XP wont even let you format an USB stick in NTFS (but only FAT). --Xerces8 09:39, 24 October 2007 (UTC)

Transparent Compression or maybe not ?

The compression is not transparent as said in the already used reference nr 16 MS KB 251186 :

Programs such as Microsoft Message Queuing (also known as MSMQ) do not work with NTFS compression.

Also see this VMware forum thread VMware eats up disk space, chkdsk needed

--Xerces8 14:41, 4 October 2007 (UTC)

Use in USB flash drive

Is it possible to format a USB flash drive with NTFS? If so, how? Stayman Apple (talk) 20:40, 19 November 2007 (UTC)

This is a discussion page on how to improve the article, not a tech support page. --Yamla (talk) 20:42, 19 November 2007 (UTC)

Well, if it isn't possible, we probably should mention that in the article. Stayman Apple (talk) 20:53, 19 November 2007 (UTC)

It's possible. Search google. --Yamla (talk) 20:54, 19 November 2007 (UTC)

successfully tried one on a 4-GB USB flash drive. Chitetskoy (talk) 14:57, 28 February 2010 (UTC)

Meta Files

I don't know much about NTFS but I'm sure that the last entry in the metafiles table is incorrect "pagefile.sys beginning of file entries." pagefile.sys is the virtual memory swap file. Kahurangikea (talk) 21:05, 30 December 2007 (UTC)

You're right, it's not a metafile. It's an "ordinary" file albeit, handled specially by the OS. I placed it there to show how the next files look. After pagefile.sys comes the \WINDOWS directory and its subdirectories. Perhaps the comment for pagefile.sys isn't pointed enough about how it's not a metafile? —EncMstr 22:46, 30 December 2007 (UTC)

Turkish I

Which file names are different: 1. FILENAME.txt; 2. FİLENAME.txt;3:25, 6 February 2008 (UTC)

Huh? #1 and #2 look identical and #3 and #4 do too. Under windows 32, these all refer to the same file. Under POSIX, they refer to two separate files. What's that have to do with Turkish? —EncMstr 17:00, 6 February 2008 (UTC)

If you look closely (and have the appropriate fonts) you'll see that two of the filenames use the Turkish dotted and dotless I. On my XP SP2 system, the two Turkish names are considered distinct from each other and from the ASCII name (which is of course case-insensitive). I don't know why. -- BenRG (talk) 19:25, 6 February 2008 (UTC)

Oh! I overlooked the dot-over-the-i differences. Thanks for pointing that out. (Perhaps I'm crippled by English?) Under win32, files 1 and 3 map to the same name; the others are distinct:

 2008-02-06  11:58                 5 FILENAME.txt
 2008-02-06  11:58                 5 FILENAME.txt
 2008-02-06  11:59                 5 Filename.txt
                3 File(s)             15 bytes

The command line dir display apparently strips the accents, but they are clearly different files. The names show up okay in a gui display. —EncMstr 20:05, 6 February 2008 (UTC)

Which file names are different under the Turkish version of Windows XP? --88.78.6.67 (talk) 15:00, 7 February 2008 (UTC)

I don't have a Turkish XP handy, though I might have Turkish NT 4.0.... The experiment above was on English XP SP2. Files 1, 2, and 4 are distinct. File 3 maps to the same name as 1. Note that is a feature of the win32 subsystem, and not of NTFS which would see all four names distinctly. —EncMstr 18:04, 7 February 2008 (UTC)

update on fat32 article required?

after reading this article, the article on fat32 seems outdated. Am I correct about this? If so, an update is desperately needed. —Preceding unsigned comment added by 145.221.52.69 (talk • contribs) 11:37, March 17, 2008

B-trees vs. B+ trees

A revert just occurred which removed a nicely cited assertion of the directory structure using B-trees. I remember reading the article when it said B+ trees and didn't think anything about it. Now that I've actually read the tree articles, I can confirm that NTFS actually uses the former. If a parent node lists a file, that filename will not (re)appear in its children. Perhaps it changed along the way? Highly unlikely methinks. —EncMstr 20:06, 16 April 2008 (UTC)

My reading of the url which mentioned B-trees is that it did not go into any technical detail, and just mentioned B-trees to give some background - useless for the purpose it was given (and as noted, it's easy enough to find articles which say it's a B+ tree). To resolve the disagreement, it's necessary to find a suitably detailed description. Tedickey (talk) 20:17, 16 April 2008 (UTC)

Just leave it as a B+ tree. Most articles about NTFS say ntfs uses B+ trees and a B+ tree is just a specialized version of a b-tree.--Thunderpenguin (talk) 21:56, 16 April 2008 (UTC)

it says b-trees on the offical microsoft website and they did create ntfs so leave it as b-tree.--24.218.246.100 (talk) 00:58, 18 April 2008 (UTC)

There you go : An MS reference that says its a B+ tree. --soum ^talk 04:37, 18 April 2008 (UTC)

So what should we do then? There is a link to b-trees and B+ trees. Maybe reading the source code of ntfs3g will help?--Thunderpenguin (talk) 23:18, 18 April 2008 (UTC)

Russinovich's reference is self-contradictory. His diagram clearly shows a B-tree, though his text calls it a B+ tree. Perhaps he was confused as to which is which. Or—maybe—the Wikipedia articles are wrong? —EncMstr (talk) 00:04, 19 April 2008 (UTC)

All B+ trees are B-trees, so it's not necessarily contradictory. It's like labelling a picture 'mammal' in one place and 'human' in another. --Yamla (talk) 00:18, 19 April 2008 (UTC)

Calling Mark Russinovich as "confused" is going a little too much. :-P As the man who is arguably the most knowledgible about the internal workings of Windows, I don't believe he is confused. Also given the number of utilities he wrote that works with internal NTFS structures, I cannot believe that they would work if they were written with the assumption NTFS uses B+ trees where they really used B-trees. Plus, do a little Google search, many sites that describe the internal structures of NTFS say its a B+ tree. Not everyone can be wrong.

Coming back to the diagram, as Yamla said, it looks like a B-tree because all B+ trees are B-trees. But that does not rule it out from being a B+ tree. There is no information to suggest that d.new, h.txt and i.doc are records. They could very well be keys for a certain record block. Consider a scenario where these files existed as records, then this structure is a valid B+ tree. Now, when those files got deleted, there is no need to delete the keys that referred to those files. B+ tree by definition allows keys to be present which are not part of the record. So, even in that case, the diagram is a valid B+ tree. --soum ^talk 05:30, 19 April 2008 (UTC)

and reading the ntfs-3g source code it says b+ trees.--Thunderpenguin (talk) 23:32, 19 April 2008 (UTC)

logfile vs journal

is a logfile the same thing as a journal? —Preceding unsigned comment added by Thunderpenguin (talk • contribs) 19:09, 12 May 2008 (UTC)

Not quite. The NTFS logfile is a journal of file system changes. It does not contain most data changes. That is, the creation of a file is logged and can be rolled back if necessary (for example, if creation of its index entry failed). Data written into the file is not normally in the log file. —EncMstr (talk) 19:32, 12 May 2008 (UTC)

Union Folders

Does this file system have union folders? If I turn on hidden files I can see two files of the same name on my desktop. On most file systems it is not possible to have two files with same name in same location. Does this mean NTFS has union folders like in Plan 9?

Bryan —Preceding unsigned comment added by 76.186.204.254 (talk) 01:58, 12 July 2008 (UTC)

No, two files cannot have the same name. They must differ in some way: different capitalization, blanks, etc. Also, Explorer combines program menus from c:/Documents and Settings/All Users/Start Menu/Programs with your individual Programs folder; it also combines Desktops in the same way. —EncMstr (talk) 06:40, 12 July 2008 (UTC)

NTFS 3.1 (XP/2003) vs NTFS 3.1 (Vista/2008)

Microsoft introduced Transactional NTFS with Vista / Windows Server 2008, but "fsutil fsinfo ntfsinfo <drive>" shows the same version?! 85.178.82.64 (talk) 08:17, 13 September 2008 (UTC)

Yep. The filesystem is exactly the same, though the filesystem driver has probably got some new stuff. Transactions are stored in an ordinary file so the filesystem doesn't have to change. —EncMstr (talk) 20:58, 13 September 2008 (UTC)

What about symbolic links wouldnt that require on disk change?--71.192.251.246 (talk) 16:46, 24 September 2008 (UTC)

No, symbolic links are implemented through "reparse points", which have always been a feature of NTFS. In fact, NTFS supported "junction points" which are disk-based directory symbolic links. The new implementation simply adds OS support for doing the same for files, as well as for allowing remote symbolic links. Ionescu007 (talk) 17:44, 16 October 2008 (UTC)

Partition Shrink Limitations

The article incorrectly states that Vista cannot move or defragment the MFT, hurting the ability to shrink the volume. This is incorrect -- the FSCTL defrag interface has supported defragging and moving the MFT since XP, and indeed the shrinking mechanism will make use of this. What can't be moved, because this is an "online" mechanism, is page file fragments and files that have been marked as "unmovable", again with a special FSCTL. An offline shrink with a 3rd-party tool would solve this, as would defragmenting the page file offline, in most cases. The website originally sourced is incorrect, because the tool they use assumes all unmovable files are MFT fragments (you can see that in the screenshot). The author then went on to assume that MFT fragments are the cause of the problem. A better defragmentation tool and disk analysis tool would reveal otherwise. Ionescu007 (talk) 17:42, 16 October 2008 (UTC)

Under XP you can't move $MFTMirr or the first 16 records of $MFT on a running system. If that's still true on Vista and if Vista puts $MFTMirr in the middle of the partition like XP does, then you won't be able to shrink a partition to less than half its original size, which is roughly what the article claimed, except it said "master file table" instead of "MFT mirror." I don't know whether any of that is true, though, as I've never used Vista. -- BenRG (talk) 18:57, 17 October 2008 (UTC)

NTFS Naming

On the original Article it is commented

"< !-- does not stand for New Technology File System: see the talk page. -- >"

I emailed sales@lsoft.net which I found on http://www.ntfs.com/quest21.htm

From: **********@******.com

Sent: October-17-08 8:28 AM
To: sales@lsoft.net
Subject: NTFS Suggestion

It is not listed on the FAQ Located at (http://www.ntfs.com/quest21.htm), I am interested to find out what does NTFS stand for?

The response I got is as follows

Hello:

New Technology File System.

Regards,

Sales Team

LSoft Technologies, Inc.

www.lsoft.net

--205.231.130.2 (talk) 15:30, 17 October 2008 (UTC)

Please see the talk archives; this issue was discussed last year and it was determined that "NT File System" and "NTFS" are the current, correct names. Briefly, the phrase "New Technology File System" was used for a while in the 1990s, but not during its inception or by the individuals who created it, and it was later dropped from all their documentation by the time Windows 2000 was released... this coincides with Microsoft dropping "New Technology" as an expansion of the letters "NT" elsewhere, too. Bill Gates himself has said that "NT" no longer has any particular meaning. Warren -talk- 15:49, 17 October 2008 (UTC)

98.242.244.0 (talk) 21:35, 15 February 2009 (UTC) I studied at one of the original Microsoft Academic Technical Education Centers, a very prestigious college, in 1998, for NT Server 4.0, we were taught at that time that NTFS, in fact, was an acronym for "New Technology File System". This was the original expansion of the acronym. Just as DOS meant "Disk Operating System". I don't care if Bill Gates himself wants to say nowadays that DOS no longer means "Disk Operating System" or if he no longer wants to admit that the original name of NTFS was "New Technology File System", it does not take away the fact that this was the original name and as such it should be included in the main article strictly for historical accuracy. Another example is the former-planet Pluto, which for all of our lives we all knew to be the 9th planet. Now they no longer want to consider it a true "planet", OK, fine, I'm sure they have their reasons, but for the sake of historical accuracy wherever there is a listing of the 8 planets, there should always remain an asterisk 9th bullet mentioning Pluto as being well known for a very long time as being a planet. Let's not rewrite history or dilute the record here, the fact is that NTFS was originally an acronym for "New Technology File System" and however this needs to be worded so it can appear on the main page, then someone please do so.

129.174.191.249 (talk) 18:39, 3 March 2009 (UTC) Modern textbooks continue to refer to NTFS as New Technology File System, so it would seem that it is still understood that the acronym represents that phrase in the field. I think it would be appropriate for at least some note, perhaps historically, to be introduced into the article.

NTFS has no specific meaning anymore. Since the merger of the Windows 9x line and NT line back in 2001, NTFS took over as the dominate file system. FAT is still around but is less used. New Technology File System is no longer accurate and hasn't been called that since the lines have merged. NT and NTFS have no specific meaning anymore. // A Raider Like Indiana 21:05, 27 April 2009 (UTC)

That might be true, however it seems that the only source for this opinion exists in this talk page. Until someone provides a reliable source, it'll be an area of dispute. Tedickey (talk) 11:19, 29 April 2009 (UTC)

Corrupted folder or file

One major problem with NTFS is that when a folder or file gets corrupted, like with a bad Internet Explorer crash (I've seen corrupted files several times in Temporary Internet Files) or some program blows up while writing to a file, there is no way I've been able to find to delete that corrupted file or folder. Windows just says it's corrupted and un-readable and won't allow anything to touch it. CHKDSK ignores them, claiming everything is 100% perfectly fine.

This is completely BACKWARDS behaviour! Corrupted files are BAD. They're USELESS. Unless you have the resources of a computer forensics lab, you're not going to get anything useful out of them even if Windows would allow the user to access them. The proper action from Windows ought to be "This file or folder is corrupted and unreadable. Do you want to delete it and recover the space?"

With prior versions of Windows and DOS running on FAT16 or FAT32, SCANDISK would 'recover' such files to useless .chk files and delete any parts it couldn't make sense of. PROBLEM SOLVED. Windows on NTFS protects corrupted files like a mother grizzly bear guarding her cubs.

Microsoft knows *everything* about how NTFS works, why haven't they provided a utility (or built into Windows) to fix this problem? Fifteen years of NTFS and it still has this problem, with no way short of reformatting the drive to fix it.

Right now I have a client's PC with XP Home that has a corrupted folder in Temporary Internet Files. It prevents the volume dirty bit from being un-set during shutdown, which causes CHKDSK to run every boot. CHKDSK reports the volume is dirty, then finds NOTHING WRONG. That user account cannot be deleted because this one folder cannot be deleted. I've tried the safe mode command prompt Administrator login to run CHKDSK /F, that doesn't do anything. I've tried it from the DaRT 5.0 CD, both connected and not to the Windows install, no luck with that either. I've also tried several utilities that are supposed to be able to forcibly delete files and folders Windows is falsely claiming are "open" or that the user "doesn't have permission" to access. None of those work.

If I had the big $$$, I'd fly to Redmond with this PC and pay Microsoft to gather their top NTFS experts together to create a utility to fix this problem- which they already should've done years ago. —Preceding unsigned comment added by 71.32.183.250 (talk) 01:49, 7 December 2008 (UTC)

This is the wrong place for this, but anyway... You keep talking about "corrupted files" that CHKDSK can't find, but how do you find them? A weird name? Weird contents (how can you tell)? Or is a "corrupted file" just one that you don't know how to delete? I'm not sure why you think this is a filesystem problem, but given that CHKDSK doesn't find anything I'm going to guess that it isn't. Your analogy with FAT doesn't make sense either—even a FAT volume can't be corrupted at the filesystem level by a crashing application. Only something that brings down the whole OS will ever leave you with a corrupted file system. Unless the driver is buggy, but that's not something that Microsoft would ignore for years. They take filesystem reliability very seriously—their corporate clients demand it.

Some practical suggestions: try taking ownership of the file/folder, try a chmod 777 using Cygwin's chmod, or try a totally non-Microsoft approach like an Ubuntu live CD. -- BenRG (talk) 14:46, 7 December 2008 (UTC)

For BenRG if you see this again, try using Ultimate Defrag a free tool which is probably the most powerful and actually opens everything on your disk so you can see it and does defrag, lines them up the way they are best, lines them up alphabetically if you want... Windows defrag doesn't work like that and in fact it's the only well thought out defrag because 90% just use windows defrag so if they dont do something special to boot they are only codding you. What the anon poster is saying is absolutely correct because I have the same thing but you'd really need to source it from some online magazine or good tech blog to add it. Maybe the magnetics in these sectors are actually damaged, try asking the ref desk... ~ R.T.G 01:30, 18 October 2009 (UTC)

Chkdsk does not check for file corruptions such as the physical data. It only checks to make sure the filesystem metadata is in tact like the master file table and folders.--Thunderpenguin (talk) 15:49, 30 May 2009 (UTC)

If you want to check for corrupted sectors, try chkdsk /l. Alternatively use scandisk and tell it to locate and fix corrupted sectors. Bear in mind this will take a long time as every sector on the disk needs to be read. Also it's not clear to me that's the problem here. Nil Einne (talk) 11:09, 25 January 2010 (UTC)

Just use a Linux LiveCD to delete the file. Simple. --Joshua Issac (talk) 18:35, 8 April 2010 (UTC)

Using Linux to modify an NTFS file system can be risky and even corrupt the file system more since NTFS on Linux hasn't been perfected. 74.218.176.50 (talk) 08:05, 24 May 2010 (UTC)

Maximum Volume Size - in normal language

In theory, the maximum NTFS volume size is 264-1 clusters. However, the maximum NTFS volume size as implemented in Windows XP Professional is 232-1 clusters. For example, using 64 kB clusters, the maximum NTFS volume size is 256 TB minus 64 kB. Using the default cluster size of 4 kB, the maximum NTFS volume size is 16 TB minus 4 kB. (Both of these are vastly higher than the 128 GB limit lifted in Windows XP SP1.) Because partition tables on master boot record (MBR) disks only support partition sizes up to 2 TB, dynamic or GPT volumes must be used to create bootable NTFS volumes over 2 TB. --IceHunter (talk) 04:01, 31 December 2009 (UTC)

WP:MOSNUM and this article

A well-intentioned editor replaced the MiBs, GiBs, etc. in this article with MB, GB, etc. citing WP:MOSNUM. However, this makes the facts of the article incorrect since powers of two truly are the proper numbers for the various limitations and structures. It's time that MOSNUM be fixed to allow for this. —EncMstr (talk) 03:35, 2 February 2010 (UTC)

WP:MOSNUM concludes the relevant section with:

The IEC standard prefixes kibi-, mebi-, gibi-, etc. (symbols Ki, Mi, Gi, etc.) are not familiar to most Wikipedia readers, so are generally not to be used except under the following circumstances:

• when the article is on a topic where the majority of cited sources use the IEC prefixes, ...

I believe this article (NTFS) clearly uses such sources, for example this one. While the article itself uses KB, MB, etc., they are clearly using those as powers of two. —EncMstr (talk) 03:44, 2 February 2010 (UTC)

The link uses KB, MB, GB, TB, EB with powers of two. The majority of cited sources in this article do not use IEC prefixes. So according to my reading of WP:MOSNUM the IEC prefixes should not be used and the first exception doesn't apply. Perhaps ask on WP:MOSNUM to double check? 203.125.91.42 (talk) 04:09, 2 February 2010 (UTC)

Supported operating systems

Recent edit added Linux, etc. Since none of those look-alike implementations provide any warranty, etc., which is associated with "support", this doesn't seem to be topical. Tedickey (talk) 09:28, 3 February 2010 (UTC)

There are two commercial companies, Tuxera and Paragon Software, which sell, support and warrant NTFS drivers for Mac OS X and Linux. —Preceding unsigned comment added by 91.153.142.162 (talk) 12:39, 3 February 2010 (UTC)

Indeed. However, beyond the mere fact of advertising its availability, those links don't add anything to the topic. Tedickey (talk) 00:27, 4 February 2010 (UTC)

Files 11 ancestry

This article has to mention that NTFS implements many concepts of the Files-11 filesystem (read the article), which is its conceptual predcessor. —Preceding unsigned comment added by 213.141.154.11 (talk) 08:12, 19 March 2010 (UTC)

For that, one needs a reliable source presenting the analysis Tedickey (talk) 12:03, 20 March 2010 (UTC)

Folders ending with a . (dot)

I use Linux as my main OS and use the NTFS-3G-driver to read and write NTFS-formatted partitions when I need to. When I create a folder ending with a . (dot) in Linux, which works fine, a computer running windows won't open them (tested with Windows 2000, Windows XP and Windows Vista). I don't see that limitation in the article. Can someone please confirm this and add it to the article? --79.212.138.66 (talk) 12:03, 14 April 2010 (UTC)

Max filename length 255 UTF-16 code units — really?

http://msdn.microsoft.com/en-us/library/aa365247.aspx#maxpath

The Windows API has many functions that also have Unicode versions to permit an extended-length path for a maximum total path length of 32,767 characters.

I decided to check the referenced source http://data.linux-ntfs.org/ntfsdoc.html.gz , but it is not available now. 83.246.135.19 (talk) 13:14, 9 November 2010 (UTC)

The terms pathname and filename are sometimes carelessly thrown about. In this case, the filename length—the number of characters for a filename (a portion of a pathname delimited by / or \)—is limited to 255 characters by the on disk storage representation (a single byte holds the string length). NTFS on disk storage does not impose any limit on a pathname. Directories can be arbitrarily nested until disk space is exhausted. Actual NTFS driver implementations also are very good about not imposing any arbitrary limits. I've created directory hierarchies under XP in which the pathnames were well over 200K. However, viewing listings of these files with Windows Explorer, etc. causes all kinds of system disruption. Also, the VC++ runtime filename parsing functions don't work too predictably—that is, some work fine, others barf internally. Visual Basic applications also tend to hold full length paths while parsing, so they aren't too good at dealing with paths over 32K. I suspect as a result of all this, Microsoft decided to deprecate paths of over 32K, but to support existing applications which work okay, they did not renege on supporting those paths within the O.S. —EncMstr (talk) 17:16, 9 November 2010 (UTC)