problems with ntfs-3g ?

BCarey · 10-29-2007, 10:42 PM

Hi,

I'll be installing a dual boot laptop soon, and would like to avoid making a fat partition if possible. ntfs-3g.org claims to have a stable read-write for ntfs partitions. Some people on this board have expressed doubts or suspicions about the reliability of ntfs-3g's product. So my question is, has anyone actually had a negative experience with this over the past, say, 6-12 months? (I'm not interested in problems older than that, please).

Thanks,
Brian

Simon Bridge · 10-29-2007, 10:45 PM

ntfs-3g is highly reliable and stable. Just a little slower than the windows native driver and it doesn't handle encrypted stuff.

Most of the people expressing doubts about ntfs-3g are recalling the high hopes and low results for the native ntfs driver.

Daedra · 10-29-2007, 11:28 PM

I have had no problems whatsoever and I have been using it for months.

bob_dobbs · 10-30-2007, 02:26 AM

It works very well lately. I have still experienced some CPU use/speed issues when writing to almost full disks, and I cant be bothered to try defragging Windows (boot windows?! - yuck!). If I get too close to filling the ntfs partition (<1GB) my CPU use climbs to max and the write speed drops to a mere 200-400KBs, which is of course no fun, but has also never failed or led to any corruption. This typically occurs with large files (>500MB).

I would not hesitate to use this driver.

brianL · 10-30-2007, 05:39 AM

No problems here with ntfs-3g. So far...fingers crossed.

vbisis · 10-30-2007, 07:28 AM

Same here, works great for me.

duryodhan · 10-30-2007, 07:56 AM

I have been using ntfs-3g with no problem for quite some time.

But the linux-ntfs site recommends to use their own fuse driver ... it seems all similar to ntfs-3g, but the linux-ntfs project is much older and more respected by me. See their wiki note on ntfs-3g.

But again, I still use 3g

shadowsnipes · 10-30-2007, 10:27 AM

NTFS-3G works great. I have experienced no noticeable slow writes/reads nor any other problems on the Linux side.

I currently still Dual-boot Windows XP and Slackware 12 so I use NTFS-3G to share NTFS partitions (versus the old way of having to use VFAT).

There is a problem on the Windows side with permissions. The NTFS-3G driver docs clearly state that they can't do perms, so be mindful of this when you create folders and files. By default (in my experience), the perms will allow Full Control to Everyone. So, on the Windows side make sure that the partition and its base folders all have correct permissions with all files/folders inheriting! Problem solved. If you add a new folder or file in Linux that won't inherit the proper permissions in Windows be sure to fix it next time you boot into Windows.

Again, this is a great tool, so anyone that needs it and hasn't tried it should be trying it.

Edit:
Use Slackbuilds to get NTFS-3G and the required FUSE library
http://slackbuilds.org/repository/12.0/system/fuse/
http://slackbuilds.org/repository/12.0/system/ntfs-3g/

enjoy!

BCarey · 10-31-2007, 10:32 PM

Thanks all for the responses. I will go ahead and use it.

Brian

syg00 · 11-01-2007, 01:26 AM

Apparently 3g is obsolete already.
I found this a few days back.
Interesting ...

duryodhan · 11-01-2007, 03:06 AM

That post is on linux-ntfs . ntfs-3g is a fork of linux-ntfs , unless ntfs-3g says it is obselete I am not going to agree with them. Especially since they have posted about that a long time back, while ntfs-3g is still releasing good drivers. Latest release was Oct 29.

I think the linux-ntfs people are just pissed off with ntfs-3g. sour grapes.

But fact is that the linux-ntfs people have many good tools (ntfsresize, mkntfs, ntfsprogs et. al) which you will require anyways, so might as well trust their judgement,

igu · 11-01-2007, 07:17 AM

ntfs-3g is a fork of linux-ntfs without the utilities by the active linux-ntfs developers. If you check who developed the most used linux-ntfs utilities then you will find that those were done by the ntfs-3g developers many years ago when they were only linux-ntfs developers. The tools weren't forked because they work fine (though it's said they're plan to fork those too since the bad quality of the linux-ntfs 2.0.0 release) but the read/write driver never worked reliable and the project leader, hired by Apple in 2005, didn't care much about it. So I also think this is a sour grapes or even worse case. ntfs-3g works for me also beautifully.

shadowsnipes · 11-01-2007, 12:04 PM

Quote:

Originally Posted by igu

So I also think this is a sour grapes or even worse case. ntfs-3g works for me also beautifully.

AGREED

Perhaps some people just don't like change... even when it is better for the whole community.

Simon Bridge · 11-01-2007, 07:00 PM

From linux-ntfs.org...

Quote:

The ntfs-3g driver is an obsolete fork of ntfsmount. Use ntfsmount from ntfsprogs-2.0.0 instead.

...wow. Talk about bad blood!

Quote:

ntfsmount (part of the ntfsprogs) is an read/write userspace NTFS driver. It provides full read-write access to NTFS, excluding writing compressed and encrypted files, changing file ownership, access right.

To be fair - ntfs-mount claims to do the same thing as ntfs-3g.

... both are FUSE (? darn! need reference), both user-space. ntfs-3g is often mistaken for a driver because of the way it is used.

Looking for comparisons:
A Slax-5.1.8 beta announcement says

Quote:

- added NTFS-3g (aka ntfsmount) to fully support writing to NTFS
partitions.

... similar statements from a range of sources show people outside the projects regarding the two as effectively identical.

I found a Linux Forum discussion (of Oct 07) covering this. Features a head-to-head between the different developers. Sample:

Yura (cha0smaster - linux-ntfs) writes:

Quote:

Wrong. ntfsprogs contain both read/write driver (ntfsmount) and tools.
Starting 2.0.0 release ntfsmount have full write support, better than 3g
speed and reading of encrypted files. So, ntfs-3g there is really no
place for ntfs-3g.

Szaka (SzabolcsSzakacsits - ntfs-3g) wrote:

Quote:

ntfsprogs 2.0.0 is so poor quality that I couldn't even test it (see e.g. kernel tarball unpacking) and finally it made me decide to integrate the ntfsprogs utilities I wrote and worked on into NTFS-3G, so there won't be need for the unreliable and inadequately maintained ntfsprogs anymore.

... and, after some users weigh in with examples of ntfsprogs poor performance, the conversation deteriorates.

Szakas main beef seems to be that 2.0.0 source was not available outside the developers - and, therefore, to him - could be sour grapes if it weren't that this strikes at the heart of normal FOSS development practice. Probably feeling left out, he seems to have gone out of his way to find fault (but still useful peer-review to do so.) Yura then feels victimized. <sigh> 'Tis always better to just point out issues - even better to provide code. This sort of back-and-forth sniping just gets boring.

There is much that is good and bad in that thread - Read it and learn.