WinFS beta out

WinFS beta out

[Gregory Maxwell]

http://it.slashdot.org/it/05/08/29/2241243.shtml?tid=109&tid=218

Looks like MSFT will be beating Linux to the nextgen FS punch after all. ;)

Re: WinFS beta out

[Jake Maciejewski]

Yeah, if SQL on NTFS counts as a revolutionary filesystem. I wonder what
Namesys could do with MS's resources.

On Tue, 2005-08-30 at 00:36 -0400, Gregory Maxwell wrote:
> http://it.slashdot.org/it/05/08/29/2241243.shtml?tid=109&tid=218
> 
> Looks like MSFT will be beating Linux to the nextgen FS punch after all. ;)
-- 
Jake Maciejewski <maciejej@msoe.edu>

Re: WinFS beta out

[Hans Reiser]

Jake Maciejewski wrote:

>Yeah, if SQL on NTFS counts as a revolutionary filesystem. I wonder what
>Namesys could do with MS's resources.
>  
>
Just a few million would be enough to do a lot.... ;-)

>On Tue, 2005-08-30 at 00:36 -0400, Gregory Maxwell wrote:
>  
>
>>http://it.slashdot.org/it/05/08/29/2241243.shtml?tid=109&tid=218
>>
>>Looks like MSFT will be beating Linux to the nextgen FS punch after all. ;)
>>    
>>

Re: WinFS beta out

[LiFe]

Watch the video demos from Ch9. Very geeky, dorky and slow.

From 10 minutes of skipping, they're basically saying they're building a 
relational database on NTFS. They're not actually rebuilding NTFS, from the 
ground up. So the sum total of the project will be a radically different 
system wide service (rich database filing), which will probably be supported 
by many applications in hereto unthought of ways.

What Hans has built is a new way of storing bits that increases speed, 
reliability and flexability in the use of those bits in hereto unthought of 
ways.

What'd be great? WinFS on Reiser4.

What we're gonna get? Windows Vista with a slowwer than ever file system, 
based on 20yo technology, and uses a ton of RAM, but where data is self 
organising in many fascinating ways. And *nix with a revolutionary fast and 
different file system, where every application is going to make different 
and incompatible use of the files as directories in order to emulate 
WinFS-like features.

http://channel9.msdn.com/ShowPost.aspx?PostID=106356#106356

Dear god I love Firefox snapback tabs. Who invents this stuff? They should 
be PAID!

----- Original Message ----- 
From: "Hans Reiser" <reiser@namesys.com>
To: "Jake Maciejewski" <maciejej@msoe.edu>
Cc: "Reiserfs List" <reiserfs-list@namesys.com>
Sent: Tuesday, August 30, 2005 4:35 PM
Subject: Re: WinFS beta out


> Jake Maciejewski wrote:
>
>>Yeah, if SQL on NTFS counts as a revolutionary filesystem. I wonder what
>>Namesys could do with MS's resources.
>>
>>
> Just a few million would be enough to do a lot.... ;-)
>
>>On Tue, 2005-08-30 at 00:36 -0400, Gregory Maxwell wrote:
>>
>>
>>>http://it.slashdot.org/it/05/08/29/2241243.shtml?tid=109&tid=218
>>>
>>>Looks like MSFT will be beating Linux to the nextgen FS punch after all. 
>>>;)
>>>
>>>
> 

Re: WinFS beta out

[Marc Perkel]

One think to remember when comparing different file systems is not just 
speed but functionality. Windows has a far better permission system than 
Linux does. ACLs bring Linux closer to Windows when it cones to fine 
grained permissions, but when you turn ACLs on with reiser 3 it slows 
down to 1/10 of normal speed and Reiser 4 ACLs just doesn't work. With 
ACLs as part of the mix ReiserFS would get its butt kicked.

Welcome to the real world.

Re: WinFS beta out

[Lexington Luthor]

Marc Perkel wrote:
> One think to remember when comparing different file systems is not just 
> speed but functionality. Windows has a far better permission system than 
> Linux does. ACLs bring Linux closer to Windows when it cones to fine 
> grained permissions, but when you turn ACLs on with reiser 3 it slows 
> down to 1/10 of normal speed and Reiser 4 ACLs just doesn't work. With 
> ACLs as part of the mix ReiserFS would get its butt kicked.
> 
> Welcome to the real world.
> 
> 

NTFS ACLs also have a number of problems.

I have worked with NT systems for many years and I would happily guve up 
that functionality for some of the flexibility of a Linux based 
infrastructure.

Sadly, at work, I am stuck with Windows :(

LL

Re: WinFS beta out

[Christian Iversen]

On Wednesday 31 August 2005 00:10, Lexington Luthor wrote:
> Marc Perkel wrote:
> > One think to remember when comparing different file systems is not just
> > speed but functionality. Windows has a far better permission system than
> > Linux does. ACLs bring Linux closer to Windows when it cones to fine
> > grained permissions, but when you turn ACLs on with reiser 3 it slows
> > down to 1/10 of normal speed and Reiser 4 ACLs just doesn't work. With
> > ACLs as part of the mix ReiserFS would get its butt kicked.
> >
> > Welcome to the real world.
>
> NTFS ACLs also have a number of problems.
>
> I have worked with NT systems for many years and I would happily guve up
> that functionality for some of the flexibility of a Linux based
> infrastructure.
>
> Sadly, at work, I am stuck with Windows :(

I can agree to this. Even with the bare essentials (groups and users only, no 
ACLs), Unix permissions have one huge gain over NT ACLs: clarity.

- When you do "ls -l", you have complete and immediate view of all nodes in 
that directory. This is many times harder to do on windows.

- When you want to figure out if a given user can access a certain file, you 
can do so from the one line in the ls output. On windows you can't do this, 
barring an add-on program, unless you click your way into a deep gui, and 
this is hard to do for multiple nodes.

- When you are securing a system, you can easily tell if a user can access a 
certain file. For instance, if /foo is "drwx------ foo bar", then you know 
that all files under /foo is only available to the one user, foo. On windows, 
if user only foo can access \foo, that's _not_ a guarantee that other users 
can't access files deeper in the directory tree. If they are given permission 
to access \foo\bar, they _can_, no matter the permissions on \bar. Logical?

- Even though administrators are supposed to be "root", they are hindered by 
ACLs, even though they can always change them. This means, that for an admin 
to get file X owned by user U, he has to either 
	- Log in as U
	- Replace the file permissions onX
both of which are not easy to do. Yes there are 3rdparty programs such as 
su.exe that will help tremendously in this, and yes you can just add 
"administrators" with "full control" to all files, but it's still bad design 
IMHO. 

Who ordered the fresh rants?

-- 
Regards,
Christian Iversen

Re: WinFS beta out

[Hans Reiser]

I strongly feel that ACLs are innately superior by design to Unix
permissions.  As for the usability issues mentioned, that is mostly due
to apps not being used to them, and is a critical mass issue and UI
implementation issue, not anything more.  I just wanted to say that, so
that all will know that it is only the implementation and UI that I
object to, not the objective.  The objective is good.

Hans

Re: WinFS beta out

[Christian Iversen]

On Wednesday 31 August 2005 03:06, Hans Reiser wrote:
> I strongly feel that ACLs are innately superior by design to Unix
> permissions.  As for the usability issues mentioned, that is mostly due
> to apps not being used to them, and is a critical mass issue and UI
> implementation issue, not anything more.  I just wanted to say that, so
> that all will know that it is only the implementation and UI that I
> object to, not the objective.  The objective is good.

I mostly agree with that - I'm sorry I didn't make it very clear in the post.

-- 
Regards,
Christian Iversen

Re: WinFS beta out

[LiFe]

Interestingly WinFS will revamp permissions drastically. Any query can be 
joined with any other query of like data.

ie, you can do a calandar query for people who you've had meetings with in 
the last 30 days. Save the query. Then do a permissions query of all 
permissions for a set of directories or files for people you've had meetings 
with in the last 30 days. Then change them all in a few clicks.

The queries are designed to return the data almost immediately, so it works 
fast.

Queries can be much more complex.


----- Original Message ----- 
From: "Christian Iversen" <chrivers@iversen-net.dk>
To: <reiserfs-list@namesys.com>
Sent: Wednesday, August 31, 2005 9:00 AM
Subject: Re: WinFS beta out


> On Wednesday 31 August 2005 00:10, Lexington Luthor wrote:
>> Marc Perkel wrote:
>> > One think to remember when comparing different file systems is not just
>> > speed but functionality. Windows has a far better permission system 
>> > than
>> > Linux does. ACLs bring Linux closer to Windows when it cones to fine
>> > grained permissions, but when you turn ACLs on with reiser 3 it slows
>> > down to 1/10 of normal speed and Reiser 4 ACLs just doesn't work. With
>> > ACLs as part of the mix ReiserFS would get its butt kicked.
>> >
>> > Welcome to the real world.
>>
>> NTFS ACLs also have a number of problems.
>>
>> I have worked with NT systems for many years and I would happily guve up
>> that functionality for some of the flexibility of a Linux based
>> infrastructure.
>>
>> Sadly, at work, I am stuck with Windows :(
>
> I can agree to this. Even with the bare essentials (groups and users only, 
> no
> ACLs), Unix permissions have one huge gain over NT ACLs: clarity.
>
> - When you do "ls -l", you have complete and immediate view of all nodes 
> in
> that directory. This is many times harder to do on windows.
>
> - When you want to figure out if a given user can access a certain file, 
> you
> can do so from the one line in the ls output. On windows you can't do 
> this,
> barring an add-on program, unless you click your way into a deep gui, and
> this is hard to do for multiple nodes.
>
> - When you are securing a system, you can easily tell if a user can access 
> a
> certain file. For instance, if /foo is "drwx------ foo bar", then you know
> that all files under /foo is only available to the one user, foo. On 
> windows,
> if user only foo can access \foo, that's _not_ a guarantee that other 
> users
> can't access files deeper in the directory tree. If they are given 
> permission
> to access \foo\bar, they _can_, no matter the permissions on \bar. 
> Logical?
>
> - Even though administrators are supposed to be "root", they are hindered 
> by
> ACLs, even though they can always change them. This means, that for an 
> admin
> to get file X owned by user U, he has to either
> - Log in as U
> - Replace the file permissions onX
> both of which are not easy to do. Yes there are 3rdparty programs such as
> su.exe that will help tremendously in this, and yes you can just add
> "administrators" with "full control" to all files, but it's still bad 
> design
> IMHO.
>
> Who ordered the fresh rants?
>
> -- 
> Regards,
> Christian Iversen 

Re: WinFS beta out

[michael chang]

On 8/31/05, LiFe <LiFers@optusnet.com.au> wrote:
> Interestingly WinFS will revamp permissions drastically. Any query can be
> joined with any other query of like data.
> 
> The queries are designed to return the data almost immediately, so it works
> fast.
> 
> Queries can be much more complex.

This sounds like it's basically a souped-up version of Window's
Indexing service... and it sounds like it could take a heck of a lot
of room.

-- 
~Mike
 - Just my two cents
 - No man is an island, and no man is unable.

Re: WinFS beta out

[LiFe]

A souped up version that makes data accessable to all applications designed 
to take advantage of it. Which, if you think about it, is what MS has been 
saying they're intending to do for ages.

As for space usage, I'm sure it won't use more than Google Desktop Search 
does, indexing every bit of everything.

Speed? The SQL engine seems to be about as fast as google desktop search, so 
pretty fast. Unfortunately, as indicated before, they seem to have based the 
bit storage mechanism on NTFS, so slow writes, massive fragmentation, 
breakable jounaling.


----- Original Message ----- 
From: "michael chang" <thenewme91@gmail.com>
To: "LiFe" <LiFers@optusnet.com.au>
Cc: "Christian Iversen" <chrivers@iversen-net.dk>; 
<reiserfs-list@namesys.com>
Sent: Wednesday, August 31, 2005 11:45 PM
Subject: Re: WinFS beta out


> On 8/31/05, LiFe <LiFers@optusnet.com.au> wrote:
>> Interestingly WinFS will revamp permissions drastically. Any query can be
>> joined with any other query of like data.
>>
>> The queries are designed to return the data almost immediately, so it 
>> works
>> fast.
>>
>> Queries can be much more complex.
>
> This sounds like it's basically a souped-up version of Window's
> Indexing service... and it sounds like it could take a heck of a lot
> of room.
>
> -- 
> ~Mike
> - Just my two cents
> - No man is an island, and no man is unable.
> 

WinFS beta out

[michael chang]

On 9/1/05, LiFe <LiFers@optusnet.com.au> wrote:
> pretty fast. Unfortunately, as indicated before, they seem to have based the
> bit storage mechanism on NTFS, so slow writes, massive fragmentation,
> breakable jounaling.

Bah, in that case, I'll probably just end up sticking with
FAT/reiserfs dual-boots (switching to Reiser4 when resizers and/or
repackers and/or converters are available).  A lot of low-end machines
don't have enough hard-disk performance to handle fragmentation well.
I think fragmentation is one of the key things that makes computers
such a pain in the ass to use.

--
~Mike
 - Just my two cents
 - No man is an island, and no man is unable.

Re: WinFS beta out

[Hans Reiser]

Marc Perkel wrote:

> One think to remember when comparing different file systems is not
> just speed but functionality. Windows has a far better permission
> system than Linux does. ACLs bring Linux closer to Windows when it
> cones to fine grained permissions, but when you turn ACLs on with
> reiser 3 it slows down to 1/10 of normal speed and Reiser 4 ACLs just
> doesn't work. With ACLs as part of the mix ReiserFS would get its butt
> kicked.
>
> Welcome to the real world.
>
>
>
I didn't have any role in the reiser3 acls.  I would never have designed
them the way they got written.  If you want high performance acls,
sponsor me to supervise the work for reiser4, and they will be very high
performance.  Making them high performance in reiser4 is straightforward
and easy.

Re: WinFS beta out

[Marc Perkel]

[-- Attachment #1: Type: text/plain, Size: 1011 bytes --]



Hans Reiser wrote:

>Marc Perkel wrote:
>
>  
>
>>One think to remember when comparing different file systems is not
>>just speed but functionality. Windows has a far better permission
>>system than Linux does. ACLs bring Linux closer to Windows when it
>>cones to fine grained permissions, but when you turn ACLs on with
>>reiser 3 it slows down to 1/10 of normal speed and Reiser 4 ACLs just
>>doesn't work. With ACLs as part of the mix ReiserFS would get its butt
>>kicked.
>>
>>Welcome to the real world.
>>
>>
>>
>>    
>>
>I didn't have any role in the reiser3 acls.  I would never have designed
>them the way they got written.  If you want high performance acls,
>sponsor me to supervise the work for reiser4, and they will be very high
>performance.  Making them high performance in reiser4 is straightforward
>and easy.
>
>  
>
Well - if it's so easy then what don't you just do it?

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com


[-- Attachment #2: Type: text/html, Size: 1662 bytes --]

Re: WinFS beta out

[David Masover]

Marc Perkel wrote:
> 
> 
> Hans Reiser wrote:
> 
>>Marc Perkel wrote:
>>
>>  
>>
>>>One think to remember when comparing different file systems is not
>>>just speed but functionality. Windows has a far better permission
>>>system than Linux does. ACLs bring Linux closer to Windows when it
>>>cones to fine grained permissions, but when you turn ACLs on with
>>>reiser 3 it slows down to 1/10 of normal speed and Reiser 4 ACLs just
>>>doesn't work. With ACLs as part of the mix ReiserFS would get its butt
>>>kicked.
>>>
>>>Welcome to the real world.
>>>
>>>
>>>
>>>    
>>>
>>I didn't have any role in the reiser3 acls.  I would never have designed
>>them the way they got written.  If you want high performance acls,
>>sponsor me to supervise the work for reiser4, and they will be very high
>>performance.  Making them high performance in reiser4 is straightforward
>>and easy.
>>
>>  
>>
> Well - if it's so easy then what don't you just do it?

#1:  Relatively easy.  Most of what Namesys does every day makes

#2:  Priorities.  They don't "just do" stuff, they do what they are
being paid to do.  If you want them to just do it, just pay them, more
than they're being paid for...

Re: WinFS beta out

[Hans Reiser]

Marc Perkel wrote:

>  
>
>>  
>>
> Well - if it's so easy then what don't you just do it?

I fund the things no one else will fund out of my pocket.  The rest I
wait for sponsors for.  I can only do so much, so I try to do what no
one else will do.

Re: WinFS beta out

[michael chang]

On 8/31/05, Hans Reiser <reiser@namesys.com> wrote:
> Marc Perkel wrote:
> > Well - if it's so easy then what don't you just do it?
> 
> I fund the things no one else will fund out of my pocket.  The rest I
> wait for sponsors for.  I can only do so much, so I try to do what no
> one else will do.
> 

And if he won't say it, I'll at least say that I'm rather greatful for this. ^^

-- 
~Mike
 - Just my two cents
 - No man is an island, and no man is unable.

Fundind Reiser Development

[Marc Perkel]

[-- Attachment #1: Type: text/plain, Size: 937 bytes --]



michael chang wrote:

>On 8/31/05, Hans Reiser <reiser@namesys.com> wrote:
>  
>
>>Marc Perkel wrote:
>>    
>>
>>>Well - if it's so easy then what don't you just do it?
>>>      
>>>
>>I fund the things no one else will fund out of my pocket.  The rest I
>>wait for sponsors for.  I can only do so much, so I try to do what no
>>one else will do.
>>
>>    
>>
>
>And if he won't say it, I'll at least say that I'm rather greatful for this. ^^
>
>  
>
I'm a little confused in that most open source projects have an army of 
volunteer programmers as opposed to paid proframmers - or - they have a 
strongly interested corporate partner. Reiser 4 has extreme potantial as 
anyone can see and I would think that the "big boys" would want it - so 
- why isn't it happening?

Also - how much funding are we talking about?

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com


[-- Attachment #2: Type: text/html, Size: 1782 bytes --]

Re: Fundind Reiser Development

[Hans Reiser]

Marc Perkel wrote:

>
>
> michael chang wrote:
>
>>On 8/31/05, Hans Reiser <reiser@namesys.com> wrote:
>>  
>>
>>>Marc Perkel wrote:
>>>    
>>>
>>>>Well - if it's so easy then what don't you just do it?
>>>>      
>>>>
>>>I fund the things no one else will fund out of my pocket.  The rest I
>>>wait for sponsors for.  I can only do so much, so I try to do what no
>>>one else will do.
>>>
>>>    
>>>
>>
>>And if he won't say it, I'll at least say that I'm rather greatful for this. ^^
>>
>>  
>>
> I'm a little confused in that most open source projects have an army
> of volunteer programmers as opposed to paid proframmers - or - they
> have a strongly interested corporate partner. Reiser 4 has extreme
> potantial as anyone can see and I would think that the "big boys"
> would want it - so - why isn't it happening?
>
> Also - how much funding are we talking about?

About $30-$40k per month is what is needed.  Volunteers are great for
things like some individual plugins, but a plugin infrastructure
requires full-time professionals. 

>-- 
>Marc Perkel - marc@perkel.com
>
>Spam Filter: http://www.junkemailfilter.com
>    My Blog: http://marc.perkel.com
>  
>

Funding Reiser Development

[Marc Perkel]

[-- Attachment #1: Type: text/plain, Size: 596 bytes --]



michael chang wrote:

>On 8/31/05, Hans Reiser <reiser@namesys.com> wrote:
>  
>
>>Marc Perkel wrote:
>>    
>>
>>>Well - if it's so easy then what don't you just do it?
>>>      
>>>
>>I fund the things no one else will fund out of my pocket.  The rest I
>>wait for sponsors for.  I can only do so much, so I try to do what no
>>one else will do.
>>
>>    
>>
>
>And if he won't say it, I'll at least say that I'm rather greatful for this. ^^
>
>  
>
Yeah - totally grateful.

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com


[-- Attachment #2: Type: text/html, Size: 1437 bytes --]

Re: WinFS beta out

[Hans Reiser]

michael chang wrote:

> And if he won't say it, I'll at least say that I'm rather greatful for
> this.

Thanks.

Re: WinFS beta out

[Marc Perkel]

[-- Attachment #1: Type: text/plain, Size: 510 bytes --]

I see. For what it's worth If I had any money I'd give it to you.

Hans Reiser wrote:

>Marc Perkel wrote:
>
>  
>
>> 
>>
>>    
>>
>>> 
>>>
>>>      
>>>
>>Well - if it's so easy then what don't you just do it?
>>    
>>
>
>I fund the things no one else will fund out of my pocket.  The rest I
>wait for sponsors for.  I can only do so much, so I try to do what no
>one else will do.
>
>  
>

-- 
Marc Perkel - marc@perkel.com

Spam Filter: http://www.junkemailfilter.com
    My Blog: http://marc.perkel.com


[-- Attachment #2: Type: text/html, Size: 1260 bytes --]

Re: WinFS beta out

[Jeff Mahoney]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hans Reiser wrote:
> I didn't have any role in the reiser3 acls.  I would never have designed
> them the way they got written.  If you want high performance acls,
> sponsor me to supervise the work for reiser4, and they will be very high
> performance.  Making them high performance in reiser4 is straightforward
> and easy.

I would have also rather designed them differently, as another item
associated with the key. One of the strengths of the reiserfs design was
the btree that allows nodes of various types. ACLs should have been
implemented this way, extents could have been implemented this way.

Unfortunately, when I tried doing that, I found that the implementation
does not at all live up to the design. The implementation can handle the
four core items, and those alone. The addition of a new item type causes
the system to panic. Those issues could have been fixed, but we still
would have been left with a filesystem that would panic if mounted on an
older kernel -- since reiserfs also has no concept of feature versioning.

The current implementation is hardly what I had envisioned ACLs and
xattrs being on reiserfs -- but was the only solution we could find to
still allow the filesystem to be mounted on an earlier kernel.

An ideal solution would have been to introduce reiserfs v3.7 with a new
magic, and add feature versioning to the superblock, similar to the way
ext[23] does it. I felt that would have been met with much, much more
resistance than the implementation I ultimately presented.

- -Jeff

- --
Jeff Mahoney
SuSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFDFdRVLPWxlyuTD7IRAmPpAJ9MoA9ge/yHyM4UvMdvSU25VMvYLwCcDoWt
qxZLg1eK8TExLasF287oWJc=
=7QFY
-----END PGP SIGNATURE-----

changing label of disk

[Bedros Hanounik]

[-- Attachment #1: Type: text/plain, Size: 449 bytes --]

here's a noob question. I have a usb HD using reiserfs 3.6 

I want to change the label of the disk, so when it's automounted, the 
mounting point has some meaning such as "backup" instead of "usbdisk"

fstab-sync seems to use the label of the disk to create a mounting point.

mkfs.reiserfs -l label will overwrite the old data; is there a way to edit 
the old lable w/o reformating the disk, other than hexedit the disk?

Thanks,

-B

[-- Attachment #2: Type: text/html, Size: 512 bytes --]

Re: changing label of disk

[evilninja]

Bedros Hanounik schrieb:
> 
> mkfs.reiserfs -l label will overwrite the old data; is there a way to edit 
> the old lable w/o reformating the disk, other than hexedit the disk?

reiserfstune -l LABEL

should do the job.

(and xfs_admin for XFS, jfs_tune for JFS ;)

-- 
BOFH excuse #219:

Recursivity.  Call back if it happens again.

Re: WinFS beta out

[Hans Reiser]

Jeff Mahoney wrote:

> Hans Reiser wrote:
>
> >I didn't have any role in the reiser3 acls.  I would never have designed
> >them the way they got written.  If you want high performance acls,
> >sponsor me to supervise the work for reiser4, and they will be very high
> >performance.  Making them high performance in reiser4 is straightforward
> >and easy.
>
>
> I would have also rather designed them differently, as another item
> associated with the key. One of the strengths of the reiserfs design was
> the btree that allows nodes of various types. ACLs should have been
> implemented this way, extents could have been implemented this way.
>
> Unfortunately, when I tried doing that, I found that the implementation
> does not at all live up to the design. The implementation can handle the
> four core items, and those alone. The addition of a new item type causes
> the system to panic. Those issues could have been fixed, but we still
> would have been left with a filesystem that would panic if mounted on an
> older kernel -- since reiserfs also has no concept of feature versioning.

So you understand why reiser4 got written....;-)

>
> The current implementation is hardly what I had envisioned ACLs and
> xattrs being on reiserfs -- but was the only solution we could find to
> still allow the filesystem to be mounted on an earlier kernel.
>
> An ideal solution would have been to introduce reiserfs v3.7 with a new
> magic, and add feature versioning to the superblock, similar to the way
> ext[23] does it. I felt that would have been met with much, much more
> resistance than the implementation I ultimately presented.

I understand that you meant well, and we just honestly disagree about
when to code freeze a stable release and put all the new stuff in the
next release.  I would never release code that dramatically reduces
performance to the point of eliciting the remarks in this thread by the
users, and that too is an honest disagreement.  I was offered money by
more than one entity to do the design you did, and refused the money.

All of what you wish you could have done differently is easily done in
reiser4 because of plugins, and acls in reiser4 will happen when someone
offers money (and they will, I am sure of it).

>
> -Jeff
>
> --
> Jeff Mahoney
> SuSE Labs

Re: WinFS beta out

[Marc Perkel]

[-- Attachment #1: Type: text/plain, Size: 1753 bytes --]



Hans Reiser wrote:

>Jeff Mahoney wrote
>
>  
>
>>The current implementation is hardly what I had envisioned ACLs and
>>xattrs being on reiserfs -- but was the only solution we could find to
>>still allow the filesystem to be mounted on an earlier kernel.
>>
>>An ideal solution would have been to introduce reiserfs v3.7 with a new
>>magic, and add feature versioning to the superblock, similar to the way
>>ext[23] does it. I felt that would have been met with much, much more
>>resistance than the implementation I ultimately presented.
>>    
>>
>
>I understand that you meant well, and we just honestly disagree about
>when to code freeze a stable release and put all the new stuff in the
>next release.  I would never release code that dramatically reduces
>performance to the point of eliciting the remarks in this thread by the
>users, and that too is an honest disagreement.  I was offered money by
>more than one entity to do the design you did, and refused the money.
>
>All of what you wish you could have done differently is easily done in
>reiser4 because of plugins, and acls in reiser4 will happen when someone
>offers money (and they will, I am sure of it).
>
>  
>
I have to question the decision to freeze existing code when the new 
code isn't finished. For what it's worth I've switched to EXT3 because I 
want ACLs. I hope to come back to ReiserFS some day but right now it 
just doesn't have the features I need now. Abandoning Reiser 3 is going 
to end up losing you part of market share because EXT3 has made a lot of 
advances.

I'm still a fan and looking forward to Reiser 4. Just not sure what I 
can do to help. I'll offer you free spam filterinf Hans if that hels a 
little. I have the best spam filter on the planet.



[-- Attachment #2: Type: text/html, Size: 2164 bytes --]

Re: WinFS beta out

[Adrian Ulrich]

> If you want high performance acls,
> sponsor me to supervise the work for reiser4, and they will be very high
> performance.  Making them high performance in reiser4 is straightforward
> and easy.

How much would it cost?

 1'000$ ? 10'000$?

 -- Adrian



-- 
"Wow, I'm Dazzled!
These graphs are so pretty!"
    - CEOs Everywhere 

WinFS beta out

[michael chang]

On 8/30/05, Marc Perkel <marc@perkel.com> wrote:
> One think to remember when comparing different file systems is not just
> speed but functionality. Windows has a far better permission system than
> Linux does. ACLs bring Linux closer to Windows when it cones to fine
> grained permissions, but when you turn ACLs on with reiser 3 it slows
> down to 1/10 of normal speed and Reiser 4 ACLs just doesn't work. With
> ACLs as part of the mix ReiserFS would get its butt kicked.
>
> Welcome to the real world.

For administrators on large networks, this makes sense.

Yet, for home users, who like performance w/o the worry for ACLs
(remember FAT?) it seems to me that there is an issue here about who's
targeting whom.

WinFS on Reiser4 would be awesome for home users.  For network
admins... it just might be a big pain, I suppose, without ACLs,
anyways.

--
~Mike
 - Just my two cents
 - No man is an island, and no man is unable.