call for testers. reiserfs_file_write() implementation.

call for testers. reiserfs_file_write() implementation.

[Oleg Drokin]

Hello!

    After some time of development and internal testing, I am releasing
    this in public (obviously before pushing it to Marcelo).
    There are two patches at
    ftp://ftp.namesys.com/pub/reiserfs-for-2.4/2.4.19.pending/testing
    These implement reiserfs_file_write() support (patches are against
    2.4.20-pre2) that should speedup writes for those applications
    that actually care about st_blksize content of stat(2) output.
    (including cp and dd of course)
    Before submittig these to Marcelo we'd like to get more wide testing,
    of course. So if you can try these and report any problems, that
    would be very much appreciated.
    Also first patch considerably speedups hole creation (I remember report
    about creating 128G empty file was taking ages, this should fix this).
    
    Please note that these patches are not compatible with Chris Mason's
    datalogging patches, though we will do a merge later of course.

Bye,
    Oleg

compatibility of 2.4.20-pre2/.19.pending & data-logging ?

[Manuel Krause]

On 08/14/2002 04:49 PM, Oleg Drokin wrote:
> Hello!
> 
>     After some time of development and internal testing, I am releasing
>     this in public (obviously before pushing it to Marcelo).
>     There are two patches at
>     ftp://ftp.namesys.com/pub/reiserfs-for-2.4/2.4.19.pending/testing
>     These implement reiserfs_file_write() support (patches are against
>     2.4.20-pre2) that should speedup writes for those applications
>     that actually care about st_blksize content of stat(2) output.
>     (including cp and dd of course)
>     Before submittig these to Marcelo we'd like to get more wide testing,
>     of course. So if you can try these and report any problems, that
>     would be very much appreciated.
>     Also first patch considerably speedups hole creation (I remember report
>     about creating 128G empty file was taking ages, this should fix this).
>     
>     Please note that these patches are not compatible with Chris Mason's
>     datalogging patches, though we will do a merge later of course.
> 
> Bye,
>     Oleg
> 

Sorry, Oleg, Dieter, Chris & others,

for "re-threading",

are the introduced patches in 2.4.20-pre2 and/or 2.4.19.pending 
compatible with Chris' patches? I didn't give them or 2.4.20-pre2 a try, 
so far.

Am I right to better wait until Chris is back and updates his 
data-logging stuff for the actual pre-kernel, namesys-patches and more - 
wait for the outstanding merge?? Anyone having experiences yet?

Thank you and best regards,


Manuel

Re: compatibility of 2.4.20-pre2/.19.pending & data-logging ?

[Oleg Drokin]

Hello!

On Thu, Aug 15, 2002 at 04:46:24AM +0200, Manuel Krause wrote:

> are the introduced patches in 2.4.20-pre2 and/or 2.4.19.pending 
> compatible with Chris' patches? I didn't give them or 2.4.20-pre2 a try, 
> so far.

The core code is still compatible with what's went into 2.4.20-pre2, but options
parser code have changed.
reiserfs_file_write() inplementation is not compatible with datalogging
patches.

> Am I right to better wait until Chris is back and updates his 
> data-logging stuff for the actual pre-kernel, namesys-patches and more - 
> wait for the outstanding merge?? Anyone having experiences yet?

You won't be able to use data-logging stuff with 2.4.20-pre2, until Cris will
make a merge.

Bye,
    Oleg

Re: compatibility of 2.4.20-pre2/.19.pending & data-logging ?

[Chris Mason]

On Thu, 2002-08-15 at 00:56, Oleg Drokin wrote:

> > Am I right to better wait until Chris is back and updates his 
> > data-logging stuff for the actual pre-kernel, namesys-patches and more - 
> > wait for the outstanding merge?? Anyone having experiences yet?
> 
> You won't be able to use data-logging stuff with 2.4.20-pre2, until Cris will
> make a merge.

Right, and that probably won't happen until early next week. 

-chris

reiserfs vs NTFS for large amount of small file transactions

[darren]

Hi all,

I am doing a school project that requires me to constantly move a large
number (100,000 to 300,000) of files from 1 server to 4 desktops.

I am currently doing it in NTFS, and the result was not too good.

Last week, I did that same thing using a linux box running ext3fs, and
the result was worse!

People in the other newsgroup told me to try reiserfs and that's why I
am here.

Are there any benchmarks of reiserfs vs ntfs?

Do you all think reiserfs is suitable for my use?

Regards
darren