Re: Good, recent FS comparison?

Re: Good, recent FS comparison?

[Al Boldi]

Tyler wrote:
> Ewan Grantham wrote:
> >I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
> >prepared to follow my normal pattern of installing ext3 as the
> >filesystem. However, I saw the interview with Hans Reiser about
> >ReiserFS4, and am now wondering if reiser has really improved enough
> >to use it, or if ext3 is still the way to go?
>
> You'd be best off trying some tests of your own, using files of the size
> and quantity you expect to use on a regular basis.  I would consider
> ext3, xfs, and reiser3/4... and run some tests with them.  We've had
> really good luck using XFS on large raids, I personally had a bad
> experience with reiserfs 3, it lost data on a USB based drive, as if it
> were never even there, even after trying the recovery tools.

Don't touch anything that doesn't do ordered-mode journaling, especially if 
you use raid, unless your data-consistency requirements don't require this.

XFS is best, but does not support ordered-mode.
reiser4 is still new.
ext3 is rock-solid!

--
Al

Re: Good, recent FS comparison?

[Tyler]

Al Boldi wrote:

>Tyler wrote:
>  
>
>>Ewan Grantham wrote:
>>    
>>
>>>I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
>>>prepared to follow my normal pattern of installing ext3 as the
>>>filesystem. However, I saw the interview with Hans Reiser about
>>>ReiserFS4, and am now wondering if reiser has really improved enough
>>>to use it, or if ext3 is still the way to go?
>>>      
>>>
>>You'd be best off trying some tests of your own, using files of the size
>>and quantity you expect to use on a regular basis.  I would consider
>>ext3, xfs, and reiser3/4... and run some tests with them.  We've had
>>really good luck using XFS on large raids, I personally had a bad
>>experience with reiserfs 3, it lost data on a USB based drive, as if it
>>were never even there, even after trying the recovery tools.
>>    
>>
>
>Don't touch anything that doesn't do ordered-mode journaling, especially if 
>you use raid, unless your data-consistency requirements don't require this.
>
>XFS is best, but does not support ordered-mode.
>reiser4 is still new.
>ext3 is rock-solid!
>
>--
>Al
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>  
>
Al... you've given us some "do's" .. can you give us some "why's" to go 
along with them? :)  I would appreciate a run-down with some more 
specific info as to what/why.

Thanks,
Tyler.


-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.1/104 - Release Date: 9/16/2005

Re: Good, recent FS comparison?

[Raz Ben-Jehuda(caro)]

what is ordered mode that xfs doesn't support ?

On 9/18/05, Tyler <pml@dtbb.net> wrote:
> 
> Al Boldi wrote:
> 
> >Tyler wrote:
> >
> >
> >>Ewan Grantham wrote:
> >>
> >>
> >>>I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
> >>>prepared to follow my normal pattern of installing ext3 as the
> >>>filesystem. However, I saw the interview with Hans Reiser about
> >>>ReiserFS4, and am now wondering if reiser has really improved enough
> >>>to use it, or if ext3 is still the way to go?
> >>>
> >>>
> >>You'd be best off trying some tests of your own, using files of the size
> >>and quantity you expect to use on a regular basis.  I would consider
> >>ext3, xfs, and reiser3/4... and run some tests with them.  We've had
> >>really good luck using XFS on large raids, I personally had a bad
> >>experience with reiserfs 3, it lost data on a USB based drive, as if it
> >>were never even there, even after trying the recovery tools.
> >>
> >>
> >
> >Don't touch anything that doesn't do ordered-mode journaling, especially if
> >you use raid, unless your data-consistency requirements don't require this.
> >
> >XFS is best, but does not support ordered-mode.
> >reiser4 is still new.
> >ext3 is rock-solid!
> >
> >--
> >Al
> >
> >-
> >To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> >the body of a message to majordomo@vger.kernel.org
> >More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> >
> Al... you've given us some "do's" .. can you give us some "why's" to go
> along with them? :)  I would appreciate a run-down with some more
> specific info as to what/why.
> 
> Thanks,
> Tyler.
> 
> 
> --
> No virus found in this outgoing message.
> Checked by AVG Anti-Virus.
> Version: 7.0.344 / Virus Database: 267.11.1/104 - Release Date: 9/16/2005
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 


-- 
Raz

Re: Good, recent FS comparison?

[Al Boldi]

Tyler wrote:
> Al Boldi wrote:
> >Don't touch anything that doesn't do ordered-mode journaling, especially
> > if you use raid, unless your data-consistency requirements don't require
> > this.
>
> Al... you've given us some "do's" .. can you give us some "why's" to go
> along with them? :)  I would appreciate a run-down with some more
> specific info as to what/why.

see 'XFS corruption on power-blackout' thread

> >XFS is best, but does not support ordered-mode.
Highest performance ratio available due to low CPU usage.

> >reiser4 is still new.
New, but promising.

> >ext3 is rock-solid!
Low performance ratio due to high CPU usage.

--
Al

Re: Good, recent FS comparison?

[Matt Stegman]

This has been discussed on the mailing list before.  Quick rundown:

"Ordered mode" means that a file's metadata isn't written until after the
file's data.  On XFS, which doesn't use ordered mode, under certain
circumstances you may see files which are the correct size, with correct
times, etc, but with only null bytes for the data.

In my experience it's not that big a deal; I can almost always easily
recreate data which could be lost by being written out just before a crash
or power loss.  I put important systems on an UPS and use stable kernels,
and I've never personally seen the null data problem.

Currently, I believe only ext3 and reiserfs support ordered mode.  I'm not
sure if reiser4's journaling is ordered or not.

-- 
Matt Stegman

On Sun, 18 Sep 2005, Tyler wrote:
>
> Al Boldi wrote:
>
> >Don't touch anything that doesn't do ordered-mode journaling, especially if
> >you use raid, unless your data-consistency requirements don't require this.
> >
> >XFS is best, but does not support ordered-mode.
> >reiser4 is still new.
> >ext3 is rock-solid!
> >
>
> Al... you've given us some "do's" .. can you give us some "why's" to go
> along with them? :)  I would appreciate a run-down with some more
> specific info as to what/why.

Re: Good, recent FS comparison?

[George N. White III]

On Sun, 18 Sep 2005, Matt Stegman wrote:

> This has been discussed on the mailing list before.  Quick rundown:
>
> "Ordered mode" means that a file's metadata isn't written until after the
> file's data.  On XFS, which doesn't use ordered mode, under certain
> circumstances you may see files which are the correct size, with correct
> times, etc, but with only null bytes for the data.

Another way to look at this: in some case, e.g., transaction processing,
the priority is to make sure no data are lost.  The opposite situation is
where you have large volumes of data coming in, e.g., remote sensing,
rendering farms, big numerical simulations.  If something breaks you are
losing data until the system is back.  With XFS you have a consistent 
filesystem immediately, but you may want to look carefully at the files
being written when the problem occurred.

> In my experience it's not that big a deal; I can almost always easily
> recreate data which could be lost by being written out just before a crash
> or power loss.  I put important systems on an UPS and use stable kernels,
> and I've never personally seen the null data problem.

Clients and other data sources like satellite dishes often break at
inconvenient times.

Write a large structured file like hdf from an NFS client and pull the
network plug or turn off the client when the client job finishes but 
before the data has all been written.  You should endup with a consistent
filesystem but a sparse file.

> Currently, I believe only ext3 and reiserfs support ordered mode.  I'm not
> sure if reiser4's journaling is ordered or not.

I don't thnk you can make a blanket statement -- different horses for 
different courses.

-- 
George N. White III  <aa056@chebucto.ns.ca>

Re: Good, recent FS comparison?

[James Northrup]

late reply... but ...

this is a benchmark I performed for grins sometime about February using two 
PATA udma5 volumes as software raid0.

http://kiwi.io-informatics.com/~jnorthrup/filesystem%20benchresults.xls


the uname -a isn't anywhere to be found but it was circa 2.6.8

the script is rpesumably still useful, for an experienced data groomer.

http://kiwi.io-informatics.com/~jnorthrup/fsbench

----- Original Message ----- 
From: "Tyler" <pml@dtbb.net>
To: "Al Boldi" <a1426z@gawab.com>
Cc: <linux-fsdevel@vger.kernel.org>; "Linux RAID Mailing List" 
<linux-raid@vger.kernel.org>; <linux-xfs@oss.sgi.com>
Sent: Sunday, September 18, 2005 1:15 AM
Subject: Re: Good, recent FS comparison?


>
> Al Boldi wrote:
>
>>Tyler wrote:
>>
>>>Ewan Grantham wrote:
>>>
>>>>I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
>>>>prepared to follow my normal pattern of installing ext3 as the
>>>>filesystem. However, I saw the interview with Hans Reiser about
>>>>ReiserFS4, and am now wondering if reiser has really improved enough
>>>>to use it, or if ext3 is still the way to go?
>>>>
>>>You'd be best off trying some tests of your own, using files of the size
>>>and quantity you expect to use on a regular basis.  I would consider
>>>ext3, xfs, and reiser3/4... and run some tests with them.  We've had
>>>really good luck using XFS on large raids, I personally had a bad
>>>experience with reiserfs 3, it lost data on a USB based drive, as if it
>>>were never even there, even after trying the recovery tools.
>>>
>>
>>Don't touch anything that doesn't do ordered-mode journaling, especially 
>>if you use raid, unless your data-consistency requirements don't require 
>>this.
>>
>>XFS is best, but does not support ordered-mode.
>>reiser4 is still new.
>>ext3 is rock-solid!
>>
>>--
>>Al
>>
>>-
>>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>>the body of a message to majordomo@vger.kernel.org
>>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
> Al... you've given us some "do's" .. can you give us some "why's" to go 
> along with them? :)  I would appreciate a run-down with some more specific 
> info as to what/why.
>
> Thanks,
> Tyler.
>
>
> -- 
> No virus found in this outgoing message.
> Checked by AVG Anti-Virus.
> Version: 7.0.344 / Virus Database: 267.11.1/104 - Release Date: 9/16/2005
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" 
> in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: Good, recent FS comparison?

[Randy.Dunlap]

On Tue, 27 Dec 2005 15:33:18 -0800 James Northrup wrote:

> late reply... but ...
> 
> this is a benchmark I performed for grins sometime about February using two 
> PATA udma5 volumes as software raid0.
> 
> http://kiwi.io-informatics.com/~jnorthrup/filesystem%20benchresults.xls
> 
> 
> the uname -a isn't anywhere to be found but it was circa 2.6.8
> 
> the script is rpesumably still useful, for an experienced data groomer.
> 
> http://kiwi.io-informatics.com/~jnorthrup/fsbench

Do you have a URL that works?


> ----- Original Message ----- 
> From: "Tyler" <pml@dtbb.net>
> To: "Al Boldi" <a1426z@gawab.com>
> Cc: <linux-fsdevel@vger.kernel.org>; "Linux RAID Mailing List" 
> <linux-raid@vger.kernel.org>; <linux-xfs@oss.sgi.com>
> Sent: Sunday, September 18, 2005 1:15 AM
> Subject: Re: Good, recent FS comparison?
> 
> 
> >
> > Al Boldi wrote:
> >
> >>Tyler wrote:
> >>
> >>>Ewan Grantham wrote:
> >>>
> >>>>I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
> >>>>prepared to follow my normal pattern of installing ext3 as the
> >>>>filesystem. However, I saw the interview with Hans Reiser about
> >>>>ReiserFS4, and am now wondering if reiser has really improved enough
> >>>>to use it, or if ext3 is still the way to go?
> >>>>
> >>>You'd be best off trying some tests of your own, using files of the size
> >>>and quantity you expect to use on a regular basis.  I would consider
> >>>ext3, xfs, and reiser3/4... and run some tests with them.  We've had
> >>>really good luck using XFS on large raids, I personally had a bad
> >>>experience with reiserfs 3, it lost data on a USB based drive, as if it
> >>>were never even there, even after trying the recovery tools.
> >>>
> >>
> >>Don't touch anything that doesn't do ordered-mode journaling, especially 
> >>if you use raid, unless your data-consistency requirements don't require 
> >>this.
> >>
> >>XFS is best, but does not support ordered-mode.
> >>reiser4 is still new.
> >>ext3 is rock-solid!
> >>
> >>--
> >>Al
> >>
> >>-
> >>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> >>the body of a message to majordomo@vger.kernel.org
> >>More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> > Al... you've given us some "do's" .. can you give us some "why's" to go 
> > along with them? :)  I would appreciate a run-down with some more specific 
> > info as to what/why.
> >
> > Thanks,
> > Tyler.


---
~Randy

Re: Good, recent FS comparison?

[James Northrup]

http://ioi.homelinux.net/~jnorthrup/filesystem%20benchresults.xls
http://ioi.homelinux.net/~jnorthrup/fsbench

Sorry, apache was being too helpful here...

----- Original Message ----- 
From: "Randy.Dunlap" <rdunlap@xenotime.net>
To: "James Northrup" <jim@grrrrr.gotdns.com>
Cc: <pml@dtbb.net>; <a1426z@gawab.com>; <linux-fsdevel@vger.kernel.org>; 
<linux-raid@vger.kernel.org>; <linux-xfs@oss.sgi.com>
Sent: Tuesday, December 27, 2005 5:45 PM
Subject: Re: Good, recent FS comparison?


> On Tue, 27 Dec 2005 15:33:18 -0800 James Northrup wrote:
>
>> late reply... but ...
>>
>> this is a benchmark I performed for grins sometime about February using 
>> two
>> PATA udma5 volumes as software raid0.
>>
>> http://kiwi.io-informatics.com/~jnorthrup/filesystem%20benchresults.xls
>>
>>
>> the uname -a isn't anywhere to be found but it was circa 2.6.8
>>
>> the script is rpesumably still useful, for an experienced data groomer.
>>
>> http://kiwi.io-informatics.com/~jnorthrup/fsbench
>
> Do you have a URL that works?
>
>
>> ----- Original Message ----- 
>> From: "Tyler" <pml@dtbb.net>
>> To: "Al Boldi" <a1426z@gawab.com>
>> Cc: <linux-fsdevel@vger.kernel.org>; "Linux RAID Mailing List"
>> <linux-raid@vger.kernel.org>; <linux-xfs@oss.sgi.com>
>> Sent: Sunday, September 18, 2005 1:15 AM
>> Subject: Re: Good, recent FS comparison?
>>
>>
>> >
>> > Al Boldi wrote:
>> >
>> >>Tyler wrote:
>> >>
>> >>>Ewan Grantham wrote:
>> >>>
>> >>>>I've just setup a nice, 6-disk, USB-2 300 Gig/disk array, and was
>> >>>>prepared to follow my normal pattern of installing ext3 as the
>> >>>>filesystem. However, I saw the interview with Hans Reiser about
>> >>>>ReiserFS4, and am now wondering if reiser has really improved enough
>> >>>>to use it, or if ext3 is still the way to go?
>> >>>>
>> >>>You'd be best off trying some tests of your own, using files of the 
>> >>>size
>> >>>and quantity you expect to use on a regular basis.  I would consider
>> >>>ext3, xfs, and reiser3/4... and run some tests with them.  We've had
>> >>>really good luck using XFS on large raids, I personally had a bad
>> >>>experience with reiserfs 3, it lost data on a USB based drive, as if 
>> >>>it
>> >>>were never even there, even after trying the recovery tools.
>> >>>
>> >>
>> >>Don't touch anything that doesn't do ordered-mode journaling, 
>> >>especially
>> >>if you use raid, unless your data-consistency requirements don't 
>> >>require
>> >>this.
>> >>
>> >>XFS is best, but does not support ordered-mode.
>> >>reiser4 is still new.
>> >>ext3 is rock-solid!
>> >>
>> >>--
>> >>Al
>> >>
>> >>-
>> >>To unsubscribe from this list: send the line "unsubscribe linux-raid" 
>> >>in
>> >>the body of a message to majordomo@vger.kernel.org
>> >>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> >>
>> > Al... you've given us some "do's" .. can you give us some "why's" to go
>> > along with them? :)  I would appreciate a run-down with some more 
>> > specific
>> > info as to what/why.
>> >
>> > Thanks,
>> > Tyler.
>
>
> ---
> ~Randy
> 

Re: Good, recent FS comparison?

[Jamie Lokier]

Al Boldi wrote:
> ext3 is rock-solid!

If only.  Recently I had a system come up after a power cycle with a
directory where reading any file in that directory gives an I/O error.
The disk is fine, and it's using ext3 in ordered mode, with IDE
write-caching disabled to be sure.

So while ext3 is good, I'm not convinced it's rock solid.

-- Jamie

Re: Good, recent FS comparison?

[Al Boldi]

Jamie Lokier wrote:
> Al Boldi wrote:
> > ext3 is rock-solid!
>
> If only.  Recently I had a system come up after a power cycle with a
> directory where reading any file in that directory gives an I/O error.
> The disk is fine, and it's using ext3 in ordered mode, with IDE
> write-caching disabled to be sure.

2.4 or 2.6?

In 2.4 try a reboot and force an fsck before mounting.

--
Al

Re: Good, recent FS comparison?

[Jamie Lokier]

Al Boldi wrote:
> Jamie Lokier wrote:
> > Al Boldi wrote:
> > > ext3 is rock-solid!
> >
> > If only.  Recently I had a system come up after a power cycle with a
> > directory where reading any file in that directory gives an I/O error.
> > The disk is fine, and it's using ext3 in ordered mode, with IDE
> > write-caching disabled to be sure.
> 
> 2.4 or 2.6?
> 
> In 2.4 try a reboot and force an fsck before mounting.

2.4.26, uclinux - it's an embedded device.

Doing an fsck before mounting would be an unacceptable boot-time delay.

Why do you suggest that, specifically for 2.4?  Is there a known
problem with 2.4 and ext3?

Thanks,
-- Jamie

Re: Good, recent FS comparison?

[Al Boldi]

Jamie Lokier wrote:
> Al Boldi wrote:
> > Jamie Lokier wrote:
> > > Al Boldi wrote:
> > > > ext3 is rock-solid!
> > >
> > > If only.  Recently I had a system come up after a power cycle with a
> > > directory where reading any file in that directory gives an I/O error.
> > > The disk is fine, and it's using ext3 in ordered mode, with IDE
> > > write-caching disabled to be sure.
>
> 2.4.26, uclinux - it's an embedded device.
>
> Doing an fsck before mounting would be an unacceptable boot-time delay.
>
> Why do you suggest that, specifically for 2.4?  Is there a known
> problem with 2.4 and ext3?

I saw this at least once on 2.4.18. A forced fsck on reboot fixed it.
2.4.31 looks ok.

--
Al