PUBLIC MARKS with tags raid & linux

Our intention is not to build a fail-safe system nor a HA-Linux. We are just using RAID-1 mirrored disks to make sure that the system can be restored easily if 1 disk crashes. The following has been tested only twice (hey, we've got other things to do). It did work for us, it might work for you. It might also make your rats go berserk or just blow your mind.

A RAID device is a Redundant Array of Independent Disks. The concept was developed in 1987 at UC Berkeley and involves the creation of a virtual disk from multiple small disks in order to deliver improved performance and reliability. There are many flavors of RAID and lots of variations in how to implement it. We detail here a specific instance we use: software RAID1 using IDE disks on a Dell PowerEdge box running Debian "sarge" loaded with grub, managed by mdadm, using the ext3 journaling file system.

The system is equipped with 2 similar hard disks (they do not have to be the same but it helps if they are at least similar). The first disk is attached to the primary IDE as master, it will be referenced as /dev/hda. The second disk is attached to the secondary IDE also as master, it will be referenced as /dev/hdc.

Ce document explique comment mettre en place plusieurs types différents de RAID logiciel sur une plateforme Linux.

Comment installer du RAID0 sur Ubuntu Edgy - blog.donax.ch

Ou comment faire une grosse partition répartie sur plusieurs disques

A l"heure où les serveurs d"entreprises doivent stocker un volume croissant de données et assurer une haute disponibilité, il est nécessaire d"imaginer des systèmes de mirroring (miroir) autres que simplement sur des disques durs. La redondance de disques durs (RAID-1, RAID-5) permet déjà d"avoir une bonne résistance aux pannes d"un disque (ou plusieurs si on est en RAID-5). Cependant, si c"est la machine qui tombe (le processeur, l"alimentation, le contrôleur de disque,...), on n"a aucun moyen pour relancer le tout rapidement sans perte de données. Ayant été confronté à cette problématique, j"ai envisagé une solution de RAID-1 au travers du réseau. C"est-à-dire que sur deux serveurs, on a deux partitions (une par serveur) qui sont à tout moment une copie exacte l"une de l"autre. C"est un mirroring (miroir) de partitions à travers une interface réseau.

DRBD is a block device which is designed to build high availability clusters. This is done by mirroring a whole block device via (a dedicated) network. You could see it as a network raid-1