Raid 1 en Linux
| Posted in GNU/Linux
sábado 17 de marzo de 2007
Raid 1 en Linux
Siguiendo el artículo anterior del servidor NAS para casa, aquí cuento todo lo que hice para configurar y probar el raid 1 en debian y las conclusiones finales a las que llegué. Este pequeño tutorial lo he hecho en una máquina virtual VMware por lo cómodo que es probar e instalar todo. El disco de sistema es hda con una única partición y los discos con los que se creará el raid 1 serán hdb y hdc. Muestro en rojo lo que he introducido en los diferentes menús.
shian:~# fdisk /dev/hdb
Command (m for help): p
Disk /dev/hdb: 1073 MB, 1073741824 bytes
16 heads, 63 sectors/track, 2080 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes
Device Boot Start End Blocks Id System
Command (m for help): n
Command action
e extended
p primary partition (1-4)
p
Partition number (1-4): 1
First cylinder (1-2080, default 1):INTRO
Using default value 1
Last cylinder or +size or +sizeM or +sizeK (1-2080, default 2080):INTRO
Using default value 2080
Command (m for help): t
Selected partition 1
Hex code (type L to list codes): fd
Changed system type of partition 1 to fd (Linux raid autodetect)
Command (m for help): p
Disk /dev/hdb: 1073 MB, 1073741824 bytes
16 heads, 63 sectors/track, 2080 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes
Device Boot Start End Blocks Id System
/dev/hdb1 1 2080 1048288+ fd Linux raid autodetect
Command (m for help): w
The partition table has been altered!
Calling ioctl() to re-read partition table.
Syncing disks.
shian:~# mdadm --create /dev/md0 --verbose --level=1 --raid-devices=2 /dev/hdb1 /dev/hdc1
mdadm: size set to 1048192K
mdadm: array /dev/md0 started.
shian:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus1/target0/lun0/part1[1] ide/host0/bus0/target1/lun0/part1[0]
1048192 blocks [2/2] [UU]
[===========>.........] resync = 58.2% (611196/1048192) finish=0.0min speed=101866K/sec
unused devices:
shian:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus1/target0/lun0/part1[1] ide/host0/bus0/target1/lun0/part1[0]
1048192 blocks [2/2] [UU]
unused devices:
shian:~# mkfs.ext3 /dev/md0
mke2fs 1.37 (21-Mar-2005)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
131072 inodes, 262048 blocks
13102 blocks (5.00%) reserved for the super user
First data block=0
8 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376
Writing inode tables: done
Creating journal (4096 blocks): done
Writing superblocks and filesystem accounting information: done
This filesystem will be automatically checked every 22 mounts or
180 days, whichever comes first. Use tune2fs -c or -i to override.
shian:~# mkdir /mnt/raid
shian:~# echo "/dev/md0 /mnt/raid ext3 defaults 0 1" >> /etc/fstab
shian:~# mount /mnt/raid
shian:~# df -h /dev/md0
Filesystem Size Used Avail Use% Mounted on
/dev/md0 1008M 17M 941M 2% /mnt/raid
A partir de aquí, lo que yo hice inicialmente fue simular que un disco duro se estropeaba y al arrancar la máquina quería montar el raid sólo con el otro disco y utilizarlo normalmente. Además, después de simular con la máquina virtual que añadía un nuevo disco duro, quería añadirlo al raid para volver a tener de nuevo la redundancia. Después de leer muchos tutoriales y foros no había manera de que funcionase. Si reiniciaba la máquina sin un disco del raid, éste no se montaba y no podía acceder a los datos. Además, el dispositivo /dev/md0 no era reconocido, por lo que era como si el raid no existiese!. Finalmente, encontré en un pequeño tutorial la solución a mis problemas.
shian:/# cd /etc/mdadm
shian:/etc/mdadm# cp mdadm.conf mdadm.conf.`date +%y%m%d`
shian:/etc/mdadm# echo "DEVICE partitions" > mdadm.conf
shian:/etc/mdadm# mdadm --detail --scan >> mdadm.conf
shian:/etc/mdadm#
shian:/etc/mdadm# cat mdadm.conf
DEVICE partitions
ARRAY /dev/md0 level=raid1 num-devices=2 UUID=a48e6816:ea6e7f37:6cc50cdb:6fead399
devices=/dev/hdb1,/dev/hdc1
shian:/etc/mdadm# umount /mnt/raid
shian:/etc/mdadm# mdadm --stop /dev/md0
shian:/etc/mdadm# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
unused devices:
shian:/etc/mdadm# mdadm --assemble /dev/md0 /dev/hdb1 /dev/hdc1
mdadm: /dev/md0 has been started with 2 drives.
shian:/etc/mdadm# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus0/target1/lun0/part1[0] ide/host0/bus1/target0/lun0/part1[1]
1048192 blocks [2/2] [UU]
unused devices:
Ahora sí, vamos a probar si realmente el podemos recuperar la información y el sistema funciona correctamente en caso de caída de un dispositivo. Además, veremos cómo reemplazar el disco defectuoso y recuperar de nuevo el raid 1 con los dos discos.
shian:/# dd if=/dev/urandom of=/mnt/raid/random1 count=51200
51200+0 records in
51200+0 records out
26214400 bytes transferred in 7.829523 seconds (3348148 bytes/sec)
shian:/# cksum /mnt/raid/random1
1652310020 26214400 /mnt/raid/random1
md: bind
md: ide/host0/bus1/target0/lun0/part1's event counter: 00000006
md0: former device hdb1 is unavailable, removing from array!
md: raid1 personality registered as nr 3
md0: max total readahead window set to 124k
md0: 1 data-disks, max readahead per data-disk: 124k
raid1: device ide/host0/bus1/target0/lun0/part1 operational as mirror 1
raid1: md0, not all disks are operational -- trying to recover array
raid1: raid set md0 active with 1 out of 2 mirrors
md: updating md0 RAID superblock on device
md: ide/host0/bus1/target0/lun0/part1 [events: 00000007]<6>(write) ide/host0/bus1/target0/lun0/part1's sb offset: 1048192
md: recovery thread got woken up ...
md0: no spare disk to reconstruct array! -- continuing in degraded mode
md: recovery thread finished ...
shian:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus1/target0/lun0/part1[1]
1048192 blocks [2/1] [_U]
unused devices:
shian:~# df -h /dev/md0
Filesystem Size Used Avail Use% Mounted on
/dev/md0 1008M 42M 916M 5% /mnt/raid
shian:~# mdadm --manage /dev/md0 --fail /dev/hdb1
mdadm: set /dev/hdb1 faulty in /dev/md0
Apagamos la máquina y cambiamos el disco duro defectuoso por uno nuevo. En el caso de VMware basta con crear un nuevo dispositivo de tipo disco duro. Además, este disco duro nuevo que añadimos va a ser de mayor tamaño que el anterior. Idealmente en un raid 1 los dos discos duros deben tener el mismo tamaño, pero linux nos proporciona la suficiente flexibilidad para que esto no sea así.
shian:~# sfdisk -d /dev/hdc sfdisk /dev/hdb
Checking that no-one is using this disk right now ...
OK
Disk /dev/hdb: 4161 cylinders, 16 heads, 63 sectors/track
sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/hdb: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = sectors of 512 bytes, counting from 0
Device Boot Start End #sectors Id System
/dev/hdb1 63 2096639 2096577 fd Linux raid autodetect
/dev/hdb2 0 - 0 0 Empty
/dev/hdb3 0 - 0 0 Empty
/dev/hdb4 0 - 0 0 Empty
Warning: no primary partition is marked bootable (active)
This does not matter for LILO, but the DOS MBR will not boot this disk.
Successfully wrote the new partition table
Re-reading the partition table ...
If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
shian:~# fdisk /dev/hdb
The number of cylinders for this disk is set to 4161.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
(e.g., DOS FDISK, OS/2 FDISK)
Command (m for help): n
Command action
e extended
p primary partition (1-4)
p
Partition number (1-4): 2
First cylinder (2081-4161, default 2081):INTRO
Using default value 2081
Last cylinder or +size or +sizeM or +sizeK (2081-4161, default 4161):INTRO
Using default value 4161
Command (m for help): p
Disk /dev/hdb: 2147 MB, 2147483648 bytes
16 heads, 63 sectors/track, 4161 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes
Device Boot Start End Blocks Id System
/dev/hdb1 1 2080 1048288+ fd Linux raid autodetect
/dev/hdb2 2081 4161 1048824 83 Linux
Command (m for help): w
The partition table has been altered!
Calling ioctl() to re-read partition table.
Syncing disks.
shian:~# mkfs.ext3 /dev/hdb2
mke2fs 1.37 (21-Mar-2005)
warning: 62 blocks unused.
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
131328 inodes, 262144 blocks
13110 blocks (5.00%) reserved for the super user
First data block=0
8 block groups
32768 blocks per group, 32768 fragments per group
16416 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376
Writing inode tables: done
Creating journal (8192 blocks): done
Writing superblocks and filesystem accounting information: done
This filesystem will be automatically checked every 30 mounts or
180 days, whichever comes first. Use tune2fs -c or -i to override.
shian:~# mkdir /mnt/tmp
shian:~# mount /dev/hdb2 /mnt/tmp
shian:~# mdadm --manage /dev/md0 --add /dev/hdb1
mdadm: hot added /dev/hdb1
shian:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus0/target1/lun0/part1[2] ide/host0/bus1/target0/lun0/part1[1]
1048192 blocks [2/1] [_U]
[=======>.............] recovery = 39.5% (415488/1048192) finish=0.1min speed=69248K/sec
unused devices:
shian:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host0/bus0/target1/lun0/part1[0] ide/host0/bus1/target0/lun0/part1[1]
1048192 blocks [2/2] [UU]
unused devices:
shian:~# umount /mnt/raid
shian:~# mount /dev/hdb1 /mnt/raid
shian:~# cksum /mnt/raid/random1
1652310020 26214400 /mnt/raid/random1
shian:~# umount /mnt/raid/
shian:~# mount /mnt/raid/
shian:~# df -h /mnt/raid/
Filesystem Size Used Avail Use% Mounted on
/dev/md0 1008M 42M 916M 5% /mnt/raid
md: ide/host0/bus0/target1/lun0/part1's event counter: 0000000c
md: ide/host0/bus1/target0/lun0/part1's event counter: 0000000c
md: raid1 personality registered as nr 3
md0: max total readahead window set to 124k
md0: 1 data-disks, max readahead per data-disk: 124k
raid1: device ide/host0/bus0/target1/lun0/part1 operational as mirror 0
raid1: device ide/host0/bus1/target0/lun0/part1 operational as mirror 1
raid1: raid set md0 active with 2 out of 2 mirrors
md: updating md0 RAID superblock on device
md: ide/host0/bus0/target1/lun0/part1 [events: 0000000d]<6>(write) ide/host0/bus0/target1/lun0/part1's sb offset: 1048192
md: ide/host0/bus1/target0/lun0/part1 [events: 0000000d]<6>(write) ide/host0/bus1/target0/lun0/part1's sb offset: 1048192
Hemos visto una manera bastante sencilla y fiable de tener nuestros datos importantes a buen recaudo. No obstante este sistema raid no sirve de nada sin una buena política de backups, puesto que no protege del borrado accidental de archivos.
Después de haber probado FreeNAS y el Raid en Debian puedo sacar en claro lo siguiente:
Ahora ya sólo me queda cambiar mi actual Duron 1200Mhz (algo que voy a hacer en las próximas semanas) puesto que ya está algo viejo y cada vez lo noto más lento y utilizarlo como servidor NAS-Debian en casa.
Cuidado con los RAID 1 ya que no son 100% seguros.
Yo tuve que mandar uno de 2 discos a ONRETRIEVAL para que me recuperaran la info ya que no encontre otra solucion. Lo que si puedo destacar de dicha empresa es que hicieron un trabajo rapido y me costo mucho menos de lo que esperaba.