FAT32 or NTFS ?

[nunoleite]

Hi!

What is the best choice about the partition in a external drive to run Portable Applications? FAT32 or NTFS?
I search the forum and found nothing about this.

I had a problem. I used to have an external drive with FAT32 and i was using Windows XP. I used several computers with no problem in running my portable applications from my external drive.

One day i tried something. I installed Windows 7 in a new computer. At the same time on my machine with Windows XP, i copied my entire external drive to another with more capacity, and that new external drive was formated in NTFS.
Then i plug that new drive with all my portable applications, into the machine with Windows 7. Suddenly i found that lots of my portable applications (EssentialPIM, Stickies, Thunderbird, ...) where giving errors, and saying that there was no rights to write in that folder or database.

After some tests, i got some conclusions.

Running portable applications in an external drive, with FAT32, and because FAT32 as no security system, it should run on every Windows.
If running in a external drive with NTFS, the portable applications have no rights to write there files, because of the permissions Windows Vista and 7 are implementing, unless using "Run as Administrator".
As i have PStart to run automatically some applications, that would be a problem and not very friendly.
So the solution was to change the NTFS security permissions to Everyone to have Full Control in the whole external drive.

So, my questions are...
Is this the corrected approach to this?
What's the best choice to have on a external drive? FAT32 or NTFS?
Using NTFS wouldn't be better because of the size of the clusters?
Using NTFS would give any other kind of problems beside this?
By the other hand is the FAT32 the best choice to use in a external drive, because it is simpler?

It would be great if this where explained more in detail on the site or here in the Forum.

Thanks
Nuno Leite

[webfork]

> Is this the corrected approach to this?

You might try right-clicking on the folder, selecting properties and making sure some kind of read-only issue isn't in force. I've had problems with that in the past on XP.

> What's the best choice to have on a external drive? FAT32 or NTFS?

This is covered in other places in much greater detail, but the short version in my experience is NTFS because it gets corrupted less. The only reason you should choose FAT32 is to make it readable by non-windows systems or you occasionally remove your drive without ejecting first.

[lyx]

Lets first tackle your win7 problems. My proposal is: Don't use computers that run vista or seven. It sux. Some may say that XP is old, but that doesnt say anything about efficiency. Some may say that xp looks less fancy, but only a moron would accept fancyness if it comes bundled with much more problems. Some may say that XP is less secure, but the truth is that the biggest security issue of windows, is the person behind the monitor. Plus vista's and seven's security measures are irritating and annoying enough, that people actually get encouraged to just "click okay" even more than they were with XP (thus, actually LOWERING security). This isn't some off-topic rant against vista/seven... what i'm trying to say is: as long as the machine is running vista or seven, you'll get problems - all kinds of em. NTFS may solve one incompatibility, but there are hundreds other cases of esoteric weirdness waiting to ruin your day. There's a reason for the slow adoption of vista/seven.

As for FAT32 and NTFS themselves...

FAT32:
- Minimum overhead - you get almost the entire drive available for data.
- Compatible with lunix and i suspect also OSX (not sure about the later)
- 4GB filesize limit
- 32GB artificial partition size limit (FAT32 actually supports much much more, but MS wants to "encourage" you to use NTFS)
- slow
- no journaling. What this means is that if the machine crashes during a filewrite, that file will become corrupted. If it crashes during a directory write and similiar stuff, things may turn out bad.
- no support for stupid security metadata.
- does not support alternate datastreams. WYSIWYG.

NTFS
- High overhead - up to 20% of the memory may be lost to a bloated MFT.
- Incompatible with lunix and possibly also OSX (forget the OSS drivers - the incompetent devs of these drivers don't even understand that NTFS filenames are case insensitive (will cause filesystem corruption))
- Filesize and partition limit is way higher than you'll ever need.
- fast
- journal support. Crashes are usually recoverable.
- supports stupid security metadata.
- supports alternate datastreams so that malware can more easily hide from your view

Summary:
FAT32 is slow and less reliable, but compatible and doesn't contain junk. NTFS is fast and reliable, but is incompatible and comes with lots of junk.

[-.-]

I've used ntfs on linux/mac os... so it's compatiable at least with the newer versions so it isnt an issue.

I mainly use fat32 on flash drives because there isnt journaling to wear it down. But it's been years since that's affected the life span of the drives so this isnt a big issue either. Also you can change this under drives properties on computer manager.

I kind of like the ntfs having compression on it but I'm not sure how effective it is.

[SYSTEM]

Lets first tackle your win7 problems. My proposal is: Don't use computers that run vista or seven. It sux. Some may say that XP is old, but that doesnt say anything about efficiency.

Windows XP is the oldest still supported operating system in the world. In addition, XP has already problems with modern hard drives. See this and this.

Some may say that XP is less secure, but the truth is that the biggest security issue of windows, is the person behind the monitor. Plus vista's and seven's security measures are irritating and annoying enough, that people actually get encouraged to just "click okay" even more than they were with XP (thus, actually LOWERING security).

In fact, Windows XP can be made more secure than Vista or 7... simply by using a limited account. I do so, even though malware has never had any chance to be installed on my machine. (Once I clicked an ad by mistake and was landed to a website which attempted to scare me to download an executable. Of course, I declined to download it.) It's called layered security.

But using a limited account on XP is annoying. If I want to edit system-wide settings, I have to use fast user switching and log in as administator.

UAC is one of the best things Microsoft ever intended. (It was mostly copied from GNU/Linux, though.) UAC not only makes using admin account safer but also makes using limited account more comfortable. It is a good compromise between security and convenience. Microsoft ruined UAC in Windows 7, though: http://www.pretentiousname.com/misc/win ... list2.html

This isn't some off-topic rant against vista/seven... what i'm trying to say is: as long as the machine is running vista or seven, you'll get problems - all kinds of em. NTFS may solve one incompatibility, but there are hundreds other cases of esoteric weirdness waiting to ruin your day. There's a reason for the slow adoption of vista/seven.

Windows 7 is selling faster than any operating system before. BTW, the access right problem with NTFS is present on all operating systems of Windows NT family - not only Vista and 7.

I use Windows XP myself , but I'm going to upgrade to Windows 7 the very same day Microsoft releases Windows 7 SP1. (At the beginning of 2011, it seems.)

FAT32:
[...]
- slow

FAT32 is actually quite fast filesystem if it is either not fragmented or on a Flash-based drive. Mind you, FAT32 fragments very fast.

NTFS
- High overhead - up to 20% of the memory may be lost to a bloated MFT.

12,5 % by default. The amount of deserved disk space can be configured between 12,5 % and 50 %. In addition, files can be stored inside MFT.

- Incompatible with lunix and possibly also OSX (forget the OSS drivers - the incompetent devs of these drivers don't even understand that NTFS filenames are case insensitive (will cause filesystem corruption))

I sometimes use NTFS-3G to access my three NTFS partitions within GNU/Linux. Filesystem corruption has occurred once, but I don't know why (I wouldn't blame NTFS-3G). One partition even has compression enabled and has never corrupted.

That said, I still use FAT32 in the two partitions I access frequently within GNU/Linux, as Linux kernel has read-write FAT32 support. Built-in support in kernel is more native than third-party driver.

Summary:
FAT32 is slow and less reliable, but compatible and doesn't contain junk. NTFS is fast and reliable, but is incompatible and comes with lots of junk.

You didn't mention the best features of NTFS: hard links and reparse points. I saved nearly 200 megabytes of disk space by replacing two copies of Java Runtime Environment with junction points.

I personally recommend NTFS for partitions which

• are located in mechanical hard drives, and
• are rarely or never accessed within operating systems other than Windows.

[lyx]

Oh dear, a lemming. Okay, i'll refute this one post and them simply dont bother.

Lets first tackle your win7 problems. My proposal is: Don't use computers that run vista or seven. It sux. Some may say that XP is old, but that doesnt say anything about efficiency.

Windows XP is the oldest still supported operating system in the world. In addition, XP has already problems with modern hard drives. See this and this.

You're replying with a logical falacy immediatelly after just that falacy was pointed out? Do you even think, before replying, or are you just blindly outputting whats stored in your head? As for the harddrive. The article describes a harddrive that needs three things: A 64-bit OS, completely new PCs BIOSes and a major update to every OS. Whats specific to XP here? Well, nothing (this HDD will not work on any currently existing PC and will not work on 32bit XP/Vista/Seven). However, it may happen that MS will not release the required 64bit OS update for WinXP for strategic reasons. Thats the only thing specific to XP about it, and its a "maybe".

In any case, a user that doesn't use a HDD larger than 2 TB wont be affected anyways. Just as users running an OS that still is the most popular OS in the world (above 50% marketshare) will be fine for the next 5 years anyways, simply because with such a marketshare, no HW manufacturer can risk losing that many potential customers.

In fact, Windows XP can be made more secure than Vista or 7... simply by using a limited account. I do so, even though malware has never had any chance to be installed on my machine. (Once I clicked an ad by mistake and was landed to a website which attempted to scare me to download an executable. Of course, I declined to download it.) It's called layered security.

But using a limited account on XP is annoying. If I want to edit system-wide settings, I have to use fast user switching and log in as administator.

UAC is one of the best things Microsoft ever intended. (It was mostly copied from GNU/Linux, though.) UAC not only makes using admin account safer but also makes using limited account more comfortable. It is a good compromise between security and convenience. Microsoft ruined UAC in Windows 7, though: http://www.pretentiousname.com/misc/win ... list2.html

You again seem to not having thought about my post at all, and are just outputting whats stored in your head. For reference, here is the quote you supposedly "replied" to:

Some may say that XP is less secure, but the truth is that the biggest security issue of windows, is the person behind the monitor. Plus vista's and seven's security measures are irritating and annoying enough, that people actually get encouraged to just "click okay" even more than they were with XP (thus, actually LOWERING security).

The "no admin-account" mantra is a myth spread by linux zombies - geeks who know tech, but have little understanding about trust, psychology, interfacing and transaction-mechanics. The problem with windows security is not some stupid account-rules (point in fact, existing account based ACL models are flawed to begin with, in that user rights equal application rights - that never made any sense and still doesnt). The actual problem with windows security is a user who has no clue, does not want to take responsibility and make decisions, AND who has no interest to change this (in fact, existing "(in)security software" promotes users who are even more lazy, naive and unresponsible). Thus, your typical windows user does not want to do what's the job of a user - he wants to be replaced. A responsible, educated user who makes decisions about trust however IS NOT REPLACABLE. No hardware or software can replace that.

The second problem with windows - and linux too by the way - is that there is no efficient, simple and nonannoying mechanics to let the user define who may access what. A computer potentially has tenthousands of files and conceptual objects - no one wants to manage trust between all those things on a per-object basis. "File-based" security rules - and even account-based security rules - were invented by idealistic engineers in the stoneage, where they worked not because they're practically efficient, but instead just because there was much less complexity to manage.

Simple example: Why can i not with 1 mouseclick tell an application to have access to my documents, but not to other applications? Why is there not even a list in which i can review my past decisions? Oh, wait, i remember - my decisions ARENT EVEN SAVED, unless i go through some complicated and difficult to understand procedure to set per-file or per-directory access rights. In a security model that can efficiently manage modern practice on a computer, none of this should exist. What instead should be the case, is the computer being grouped into various "areas/categories", and then being able to say which area may interact with which area. In principle, this is a bit like how accessrules of a firewall work, except of less finegrained (not individual objects and accesses are managed, but instead higher level structures). This is NOT the same as giving an application "more" or "less" access. "More or less" is stoneage and practically inefficient - a model invented for companies and their "departments" in the early 1990s. What should happen is simply that the user permanently decides what on his computer may interact with what - i.e. "this app may access the network and my media, but not other applications nor the platform".

But: For this to work efficiently, there would need to be a user who is willing to do the job of a user. And existing organization structures of filesystems, accounts and even OS libraries would need to be different. Both are things which the market does not want to do. This is the actual problem: The market does not want efficient security - it wants a replacement for it. It wants a cheap lazy patch, so that it doesn't need to invest the effort to do it properly (you know, the usual human-style). Problem is: these cheap patches just create more problems - but you can always believe and hope.

Windows 7 is selling faster than any operating system before. BTW, the access right problem with NTFS is present on all operating systems of Windows NT family - not only Vista and 7.

There are two problems here: First, you don't verify the arguments of your sources much, dont you? Here are some hints:

A) A sold copy tells you nothing about the market impact, unless you also take the SIZE of the market into account. Pure coincidence: Today, there are also more computers than ever. OF COURSE IT SELLS MORE COPIES THAN ANYTHING BEFORE. Blergh. That number by itself alone is pure marketing blahblah. It only becomes a useful info, when its related to the number of computers in use.

B) XP right now has over 50% marketshare. An OS typically has a "boom"-like curve - at the begin, the most copies are sold. However, even if we assume it were linear, the problem is that seven sells "so fast", that it will take 4 years until XP's and Vistas marketshare has gone down enough that they dont matter much anymore. But as described before, the process will slow down, so that it will actually take longer than that. XP is here to stay for at minimum another 5 years, and no rethorics and wishful thinking on your side will change that.

I'm bored by now and dont want to reply to the rest of your post, because most arguments are based on false or missing information, and logical falacies.

[SYSTEM]

Oh dear, a lemming. Okay, i'll refute this one post and them simply dont bother.

OK, I don't expect a reply to this post.

Lets first tackle your win7 problems. My proposal is: Don't use computers that run vista or seven. It sux. Some may say that XP is old, but that doesnt say anything about efficiency.

Windows XP is the oldest still supported operating system in the world. In addition, XP has already problems with modern hard drives. See this and this.

You're replying with a logical falacy immediatelly after just that falacy was pointed out? Do you even think, before replying, or are you just blindly outputting whats stored in your head?

If supporting an operating system over a decade is a good idea, why no one has done it before? (Windows 98 was supported eight years - XP will reach 10 years in August 2011. The record of a GNU/Linux distribution is seven years: http://www.redhat.com/security/updates/errata/)

As for the harddrive. The article describes a harddrive that needs three things: A 64-bit OS, completely new PCs BIOSes and a major update to every OS. Whats specific to XP here?

First of all, new BIOSes are only required in order to boot from a hard drive whose logical sector size is four kilobytes. Most problems can be worked around by tricking the computer to believe that the sector size is still 512 bytes.

But there is one problem caused by XP:

It would appear that Windows XP not only expects misaligned partitions; it actually will not function properly without them. One simply cannot run XP on a device which has been properly partitioned for 4K physical sector sizes.

If the first partition begins in (logical) sector 64, Windows XP will not boot. If it begins in sector 63, XP will boot, but every write operation requires two read-modify-write cycles, decreasing random write performance to awkward 480 kilobytes per second (assuming 7200 RPM drive).

Windows XP is the newest operating system with this problem. Both GNU/Linux and Windows Vista/7 are able to boot even if the first partition begins in sector 64.

To cope with that, drive manufacturers have introduced an even worse hack: shifting all 512-byte logical sectors forward by one, so that logical sector 64 lands at the beginning of a physical sector. So any partitioning tool which wants to lay things out properly must know where the origin of the device actually is - and not all devices are entirely forthcoming with that information.

This makes the problem even worse. Partitioning tools don't know if the logical sectors have been shifted. If the tool guesses wrong, random write performance will be awkward, even if Windows XP is not installed.

As for the three-terabyte drives...

One of the required improvements is GUID Partition Table. GPT allows partitions to be over two terabytes in size and begin in locations after the first 2 TB.

All versions of Windows are unable to boot from a GPT disk on a BIOS-based computer. However, Windows Vista and 7, including 32-bit versions, are able to read GUID Partition Tables and see partitions in the disk. The 32-bit version of Windows XP is unable to do so.

After a couple of years, portable hard drives whose sizes exceed two terabytes will be common. The owners of such drives have to choose between XP compatibility (old MBR partition table) and full capacity of the drive (GPT).

In any case, a user that doesn't use a HDD larger than 2 TB wont be affected anyways. Just as users running an OS that still is the most popular OS in the world (above 50% marketshare) will be fine for the next 5 years anyways, simply because with such a marketshare, no HW manufacturer can risk losing that many potential customers.

This is a part of the problem. Hard drive manufacturers are already introducing hacks in order to support Windows XP better... causing problems to everyone else.

In fact, Windows XP can be made more secure than Vista or 7... simply by using a limited account. I do so, even though malware has never had any chance to be installed on my machine. (Once I clicked an ad by mistake and was landed to a website which attempted to scare me to download an executable. Of course, I declined to download it.) It's called layered security.

But using a limited account on XP is annoying. If I want to edit system-wide settings, I have to use fast user switching and log in as administator.

UAC is one of the best things Microsoft ever intended. (It was mostly copied from GNU/Linux, though.) UAC not only makes using admin account safer but also makes using limited account more comfortable. It is a good compromise between security and convenience. Microsoft ruined UAC in Windows 7, though: http://www.pretentiousname.com/misc/win ... list2.html

You again seem to not having thought about my post at all, and are just outputting whats stored in your head. For reference, here is the quote you supposedly "replied" to:

Some may say that XP is less secure, but the truth is that the biggest security issue of windows, is the person behind the monitor. Plus vista's and seven's security measures are irritating and annoying enough, that people actually get encouraged to just "click okay" even more than they were with XP (thus, actually LOWERING security).

The "no admin-account" mantra is a myth spread by linux zombies - geeks who know tech, but have little understanding about trust, psychology, interfacing and transaction-mechanics. The problem with windows security is not some stupid account-rules (point in fact, existing account based ACL models are flawed to begin with, in that user rights equal application rights - that never made any sense and still doesnt). The actual problem with windows security is a user who has no clue, does not want to take responsibility and make decisions, AND who has no interest to change this (in fact, existing "(in)security software" promotes users who are even more lazy, naive and unresponsible). Thus, your typical windows user does not want to do what's the job of a user - he wants to be replaced. A responsible, educated user who makes decisions about trust however IS NOT REPLACABLE. No hardware or software can replace that.

The second problem with windows - and linux too by the way - is that there is no efficient, simple and nonannoying mechanics to let the user define who may access what. A computer potentially has tenthousands of files and conceptual objects - no one wants to manage trust between all those things on a per-object basis. "File-based" security rules - and even account-based security rules - were invented by idealistic engineers in the stoneage, where they worked not because they're practically efficient, but instead just because there was much less complexity to manage.

Simple example: Why can i not with 1 mouseclick tell an application to have access to my documents, but not to other applications? Why is there not even a list in which i can review my past decisions? Oh, wait, i remember - my decisions ARENT EVEN SAVED, unless i go through some complicated and difficult to understand procedure to set per-file or per-directory access rights. In a security model that can efficiently manage modern practice on a computer, none of this should exist. What instead should be the case, is the computer being grouped into various "areas/categories", and then being able to say which area may interact with which area. In principle, this is a bit like how accessrules of a firewall work, except of less finegrained (not individual objects and accesses are managed, but instead higher level structures). This is NOT the same as giving an application "more" or "less" access. "More or less" is stoneage and practically inefficient - a model invented for companies and their "departments" in the early 1990s. What should happen is simply that the user permanently decides what on his computer may interact with what - i.e. "this app may access the network and my media, but not other applications nor the platform".

But: For this to work efficiently, there would need to be a user who is willing to do the job of a user. And existing organization structures of filesystems, accounts and even OS libraries would need to be different. Both are things which the market does not want to do. This is the actual problem: The market does not want efficient security - it wants a replacement for it. It wants a cheap lazy patch, so that it doesn't need to invest the effort to do it properly (you know, the usual human-style). Problem is: these cheap patches just create more problems - but you can always believe and hope.

OK, I agree. In any case, UAC is an improvement. UAC can be disabled or configured to elevate all applications without prompting if you don't like it.

Windows 7 is selling faster than any operating system before. BTW, the access right problem with NTFS is present on all operating systems of Windows NT family - not only Vista and 7.

There are two problems here: First, you don't verify the arguments of your sources much, dont you? Here are some hints:

A) A sold copy tells you nothing about the market impact, unless you also take the SIZE of the market into account. Pure coincidence: Today, there are also more computers than ever. OF COURSE IT SELLS MORE COPIES THAN ANYTHING BEFORE. Blergh. That number by itself alone is pure marketing blahblah. It only becomes a useful info, when its related to the number of computers in use.

You wrote: "There's a reason for the slow adoption of vista/seven." I still think the adoption of Windows 7 is not slow. Win7 has overtook Vista already.

B) XP right now has over 50% marketshare. An OS typically has a "boom"-like curve - at the begin, the most copies are sold. However, even if we assume it were linear, the problem is that seven sells "so fast", that it will take 4 years until XP's and Vistas marketshare has gone down enough that they dont matter much anymore. But as described before, the process will slow down, so that it will actually take longer than that. XP is here to stay for at minimum another 5 years, and no rethorics and wishful thinking on your side will change that.

You're right, and I haven't claimed otherwise.

I'm bored by now and dont want to reply to the rest of your post, because most arguments are based on false or missing information, and logical falacies.

Sad. One of the reasons I like arguing with other people is that I want to learn things myself, challenge my opinions, find out if I am right in the end.

If I'm not told what exactly is wrong, my only option is to keep my current opinions.

But feel free to not respond if you wish.

[Hydaral]

This isn't some off-topic rant against vista/seven... what i'm trying to say is: as long as the machine is running vista or seven, you'll get problems - all kinds of em. NTFS may solve one incompatibility, but there are hundreds other cases of esoteric weirdness waiting to ruin your day. There's a reason for the slow adoption of vista/seven.

Windows 7 is selling faster than any operating system before. BTW, the access right problem with NTFS is present on all operating systems of Windows NT family - not only Vista and 7.

Heh. Just because something is popular, doesn't mean it is good, or even works properly. The general public respond remarkably well to marketing, especially the ignorant ones when it comes to IT matters. There are many examples which I will not mention, this topic already seems to be a little hostile.

[SYSTEM]

Heh. Just because something is popular, doesn't mean it is good, or even works properly. The general public respond remarkably well to marketing, especially the ignorant ones when it comes to IT matters. There are many examples which I will not mention, this topic already seems to be a little hostile.

True. Maybe I should have mentioned this myself (Windows XP is absolutely very popular).