writebarries JFS's en 2.6.3x inestables / JFS's unstable on 2.6.34 and above

kernels 2.6.34 + en adelante solo soportan ext2, ext3 y ext4 establemente, el resto puede fallar y JFS falla terriblemente

kernel 2.6.34 and above only supports stability on ext2 etx3 and ext4 filesystems, other are not very stables and JFS are unstable definitiv

http://shutendoujinet.wordpress.com/2010/11/19/jfs-on-kernel-2-6-342-6-35-and-above-unstable/

http://shutendouji-net.blogspot.com/2010/11/jfs-on-kernel-2634-and-above-unstable.html

El prolema aún ocurre, esta desde el kernel "2.6.33" y aparece debido al reciente retorno de "writebarriers" en los sistemas de ficheros con journaling, en ext3 es muy estble y en ext4 se corrigieron muchas fallas pero el problema aun ocurre, aqui un hilo de la rama "2.6.34" aun sin resolver en kernel.org y debian aun no adopta los kernels superiores debido a que probando la version "2.6.35" el problema aun persiste

The prolem still occurs, from the kernel "2.6.33" and appears due to the recent return of "writebarriers" in journaling file systems like ext3 that are very stable and  about ext4 corrected many bugs, but the problem still occurs, here a thread of the branch "2.6.34" and currently debian not moved these version due testing the version "2.6.35" the problem still persists

BUG in kernel's org mailing list:

https://bugzilla.kernel.org/show_bug.cgi?id=9554 (introduction of writebarriers)

https://bugzilla.kernel.org/show_bug.cgi?id=15906 (still open, asocied to ext4, bug affects JFS too)

El primero introduce write barriers en 2.6.33 al 2.6.34 y el segundo muestra las consecuencias aun sin resolver!!!

Claramente esto deja a ext3 como uno de los mas rapidos y mejores sistemas de ficheros pero la sorpresa es que reiserfs4 no es afectado, ya que no implementa esto de la misma forma, fue un error no aceptarlo como sistema de ficheros experimental!!!

Clearly this leaves ext3 as one of the fastest and best file systems but the surprise is that reiserfs4 is not affected, due that does not implement this in the same way, it was a mistake not to accept it as experimental filesystem in kernel main line!!!