Remove NTFS kernel support

On Thursday 07 February 2008 13:21:34 Joao Barros wrote:
> On Feb 7, 2008 1:00 AM, Attilio Rao <attiliofreebsd.org> wrote:
> > As exposed by several users, NTFS seems to be
broken even before first
> > VFS commits happeing around the end of December.
Those commits exposed
> > some problems about NTFS which are currently under
investigation.
> > Ultimately, This filesystem is also unmaintained
at the moment.
> >
> > Speaking with jeff, we agreed on what can be a
possible compromise:
> > remove the kernel support for NTFS and maybe take
care of the FUSE
> > implementation.
> > What I now propose is a small survey which can
shade a light on us
> > about what do you think about this idea and its
implications:
> > - Do you use NTFS?
>
> Yes. Important in a dual booting enviroment.
>
> > - Are you interested in maintaining it?
>
> I would If I had the needed knowledge in FS and Kernel.
I only have
> availability to offer.
>
> > - Do you know a good reason to not use FUSE ntfs
implementation? What
> > the kernel counter part adds?
>
> Yes: Speed.

I think this is related only for FreeBSD:
http://www.nt
fs-3g.org/performance.html
I've used it on linux before and the performance was in pair
to kernel 
implementation. 

> A year ago when building my zfs box I had to migrate
500GB of data off
> NTFS. FUSE ntfs is WAY slow. I didn't do a proper
benchmark then but I
> could setup something now if interest arises.
> I didn't have any problems like those being reported
with CURRENT from
> April 2007 if I recall the date correctly when I copied
all that data.
>
> > - Do you think axing the kernel support a good
idea?
>
> Yes if...
>
> Yes if FUSE ntfs can have performance on par with the
current ntfs
> support. Yes if FUSE ntfs license model doesn't become
an issue.
> Yes because FUSE ntfs write support is neat =)



-- 
============================================================
==========  
- Best regards, Nikolay Pavlov.
<<<-----------------------------------    
============================================================
==========  

_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

On Thursday 07 February 2008 14:47:41 Eric Anderson wrote:
> FUSE is slow, requires a port (unless PUFFS is ported,
which I've probed
> about before).

I think this is not an argument:
http://www.nt
fs-3g.org/performance.html

-- 
============================================================
==========  
- Best regards, Nikolay Pavlov.
<<<-----------------------------------    
============================================================
==========  

_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

On Feb 7, 2008 5:20 PM, Robert Watson <rwatsonfreebsd.org> wrote:
> On Thu, 7 Feb 2008, Attilio Rao wrote:
>
> > 2008/2/7, Andre Oppermann <andrefreebsd.org>:
> >
> >> Eric Anderson wrote:
> >>> I think Alfred's point is really
interesting.  How many people that don't
> >>> use it that say 'axe it' does it take to
override 1 person saying 'keep
> >>> it!'?
> >>
> >> The real question is how many people does it
take to say 'I'll maintain
> >> it'?  Just one.  Without it, it will only
bitrot as evidenced by Attilios
> >> question.  NTFS is currently broken, just not
as obvious because WITNESS
> >> didn't track and enforce lockmgr locks.
> >
> > Andre catched exactly my point. The big problem is
that we have a list of
> > several unmaintained fs. NTFS is in this list. The
support is not reliable,
> > it is only available in read mode and eventually
bugged. I'm not sure I want
> > to keep this if nobody wants to maintain it.
>
> If you axe write support, does the maintainability of
the kernel ntfs get
> easier?  As I understand it, the write support is
rather limited, and

If I recall correctly ntfs volumes are mounted readonly by
default
(I'm unable to verify now).

> debugging and fixing read support is generally a lot
easier for a variety of
> reasons.  There's also a lot less risk to data.   I think
it's reasonable
> to surmise that, given our rather limited write support
currently, the kernel
> ntfs code is used for data migration and limited
sharing to FreeBSD in various
> forms, but that msdofs remains the general data
transport of choice...

With this in mind, I used FAT32, but occasional Windows
crashes lead
to some filesystem corruption and time consuming fsck. I
converted the
fs to ntfs and had no more issues.
General data transport of choice for usb pens, external
disks, iPods,
and when you need the ability to read/write it everywhere,
but for
running Windows it's not the best choice when compared to
ntfs.

If you think of it, FAT32 is the best supported (r/w) fs (on
disk) by
all platforms: Windows, FreeBSD, OS X, Linux.
Having read the news today about the corporate support
OpenID got, I
dream of a *good* universally supported fs. But I digress
=)

>
> Robert N M Watson
> Computer Laboratory
> University of Cambridge

-- 
Joao Barros
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

On Feb 7, 2008 5:41 PM, Nikolay Pavlov <qpadlagmail.com> wrote:
> On Thursday 07 February 2008 14:47:41 Eric Anderson
wrote:
> > FUSE is slow, requires a port (unless PUFFS is
ported, which I've probed
> > about before).
>
> I think this is not an argument:
> http://www.nt
fs-3g.org/performance.html
>

Eric has valid points.
How relevant is a benchmark on Linux to your argument?

-- 
Joao Barros
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

On Thursday 07 February 2008 20:18:42 Joao Barros wrote:
> On Feb 7, 2008 5:41 PM, Nikolay Pavlov <qpadlagmail.com> wrote:
> > On Thursday 07 February 2008 14:47:41 Eric
Anderson wrote:
> > > FUSE is slow, requires a port (unless PUFFS
is ported, which I've
> > > probed about before).
> >
> > I think this is not an argument:
> > http://www.nt
fs-3g.org/performance.html
>
> Eric has valid points.
> How relevant is a benchmark on Linux to your argument?

But it's a userland application. This page is demonstration
of it's 
potential performance that could be achieved, but were is
the FreeBSD NTFS 
implementation stats? Let me ask you: compered to what FUSE
is slow?     

-- 
============================================================
==========  
- Best regards, Nikolay Pavlov.
<<<-----------------------------------    
============================================================
==========  

_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

Robert Watson wrote:

> If you axe write support, does the maintainability of
the kernel ntfs
>  get easier?  As I understand it, the write support is
rather
> limited, and debugging and fixing read support is
generally a lot
> easier for a variety of reasons.

The few serious attempts I've made to use the NTFS in the
base to write
data were unsuccessful, and given the instability of the
code I didn't
push it very hard since I didn't want to scramble my data.

> There's also a lot less risk to data. 

Right. 

> I think it's reasonable to surmise that, given our
rather limited
> write support currently, the kernel ntfs code is used
for data
> migration and limited sharing to FreeBSD in various
forms, but that
> msdofs remains the general data transport of choice...

Read support that works is better than R/W support that
doesn't, yes. So 
if someone wants to step up to maintain this and axing write
support 
gets us something that actually works, who am I to argue?

I think keeping the goal(s) of R/W support in the base,
and/or working 
FUSE in mind is a Good Thing, but getting a working solution
to at least 
one of these problems would be a big step forward.

Doug

-- 

     This .signature sanitized for your protection
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

In message:
<3bbf2fe10802061700p253e68b8s704deb3e5e4ad086mail.gmail.com>
            "Attilio Rao" <attiliofreebsd.org> writes:
: - Do you use NTFS?
: - Are you interested in maintaining it?
: - Do you know a good reason to not use FUSE ntfs
implementation? What
: the kernel counter part adds?
: - Do you think axing the kernel support a good idea?

I use NTFS extensively and it has worked as recently as
early in the
7.0 release branch.  I think axing it is a horrible idea.

Warner
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

In message: <47AABEEF.7020807FreeBSD.org>
            Doug Barton <dougbfreebsd.org> writes:
: Alfred Perlstein wrote:
: 
: > Maybe a nicer way of saying/asking would be to ask:
: > 
: > Is the FUSE replacement going to be tested to the
point where it's
: > better than then current NTFS code?
: 
: Given that the current NTFS code in the base panics within
minutes of 
: any kind of serious access, and has the ability to take
the other 
: filesystems down with it (including UFS2) that won't be
hard.

This change in behavior is very recent.  It used to be the
one file
system you could count on to get data off a disk that was
throwing
disk errors back at the OS.

Warner
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

On Wed, Feb 06, 2008 at 11:13:14PM -0800, Alfred Perlstein
wrote:
> * Attilio Rao <attiliofreebsd.org> [080206
17:00] wrote:
> > remove the kernel support for NTFS and maybe take
care of the FUSE
> > implementation.
..
> 
> My pragmatic view on this is that I think it's odd that
something
> that is sort-of working for a few people is going to be
axed by
> people that don't use it, while promising to replace it
with something
> better.
> 
> Maybe a nicer way of saying/asking would be to ask:
> 
> Is the FUSE replacement going to be tested to the point
where it's
> better than then current NTFS code?

Hear, Hear!

_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"

Attilio, good day.

Thu, Feb 07, 2008 at 02:00:41AM +0100, Attilio Rao wrote:
> As exposed by several users, NTFS seems to be broken
even before first
> VFS commits happeing around the end of December. Those
commits exposed
> some problems about NTFS which are currently under
investigation.
> Ultimately, This filesystem is also unmaintained at the
moment.
> 
> Speaking with jeff, we agreed on what can be a possible
compromise:
> remove the kernel support for NTFS and maybe take care
of the FUSE
> implementation.
> What I now propose is a small survey which can shade a
light on us
> about what do you think about this idea and its
implications:
> - Do you use NTFS?

Yes, especially on the multi-homed notebook systems.  In
read-only
mode it rocks.

> - Are you interested in maintaining it?

Yes.  If you can throw the buglist for NTFS on me, I will be
very
grateful.

> - Do you know a good reason to not use FUSE ntfs
implementation? What
> the kernel counter part adds?

Don't know, newer tried FUSE.

> - Do you think axing the kernel support a good idea?

IMO, not a good one.
-- 
Eygene
_______________________________________________
freebsd-fsfreebsd.org mailing list

http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to
"freebsd-fs-unsubscribefreebsd.org"