Filesystems

De gacq wiki
Revisión del 16:51 27 jul 2006 de 192.168.0.105 (discusión)
(dif) ← Revisión anterior | Revisión actual (dif) | Revisión siguiente → (dif)
Saltar a: navegación, buscar

Filesystem Hierarchy Standard


The one important thing to remember is that what ever file system you are using on your root partion must be compiled into the kernel; it cannot be a module! This is not 100% true since there are ways round it but you are asking for trouble if the kernel you are using can't read your root file system.

reiserfs vs ext3: Algunas frases sueltas

The conclusion is obvious by the "Total Time For All Benchmarks Test." The best journaling file system to choose based upon these results would be: JFS, ReiserFS or XFS depending on your needs and what types of files you are dealing with. I was quite surprised how slow ext3 was overall, as many distributions use this file system as their default file system.

Basically, reiserFS is much faster on a filesystem with small-size files, and utilization of space is much better. With ext3 you aren't likely to get corrupted files in a case of unattended reboot (when you pull a plug for instance) due to a different "flushing" technique. JFS is highly recommended for enterprises, and XFS is a new kid in town that supports large files (therabytes of space).

ext3 certainly has its advantages. Being able to mount it ext2 is really nice if you accidently forget to compile ext3 support into your kernel. The ability to convert form ext2 to ext3 without redoing the partition or loosing data is pretty nice as well.

I've been using ext3 for a while, and am not particularly happy with it. It seems to cause some pretty bad performance hits (especially with tar)

GACQ: En una particion de 100 GB pasando de ext3 a reiserfs 3.6 logre una ganancia de espacio (informada por el df) de 1.57 GBs y un consumo de disco para el mismo conjunto de archivos de 180 MB menos en aproximadamente 50 GB de archivos.

Enlaces

Algunos comandos de administracion

# Ver los filesystems soportados
cat /proc/filesystems

# Ver los parametros de un filesystem
dumpe2fs -h /dev/hda1

If you're a production sysadmin and want to have better control over when the consistency check occurs, you can disable the automatic check by running the following command (replacing the device file name with the appropriate one):

tune2fs -i 0 -c 0 /dev/sdc1

Configurando una particion ReiserFS en debian sarge

Segun legue a entender tenemos 2 opciones principales:

  • Usar ReiserFS4: esto implica patchear el kernel, todavia no muy estable
  • Usar Reiserfs 3.x que esta soportado en kernels 2.4/6.xx, donde xx es algo medianamente nuevo.

Para la segunda opcion tenemos que tener el paquete con los utilitarios, para esto encontramos dos paquetes con nombre similar lo que puede causar alguna confusion:

  • reiserfsprogs
    • this package NOT use libreiserfs to do its work?
    • El paquete reiserfsprogs es de www.namesys.com
  • progsreiserfs
    • this package use libreiserfs to do its work,
modprobe reiserfs
apt-get install reiserfsprogs
mkfs -t reiserfs /dev/hdc2

Tunning del filesystem

Particiones muy grandes

ext3

  • Bajar cantidad de inodos, podemos hacerlo con la opcion -T largefile
  • Bajar el porcentaje reservado del % 5 al % 1 si no vamos a poner una aplicacion critica
mkfs.ext3 -T largefile -m 1 /dev/hdc1