Joao Rochate <[EMAIL PROTECTED]> writes:
> * If I have a software RAID enabled system and my server fails, I can move
> all the disks to another server and use the same /etc/raidtab to get disks
> working again.
Yes. If you use current alpha snapshot code, persistent superblocks
and array autodetection, you won't even need to care about the disk
order, the system will figure it out.
> * If I have a hardware RAID system and server fails (also the RAID
> controler) can I just move the disks to another server with the same RAID
> card model and have the system working back again?
Should work, although a controller might have some settings stored in
NVRAM instead of on the disks (would seem stupid to me, but..)
> I'm trying to do this cheap near-HA solution working. The software RAID
> takes too much time to rebuild a failed 40Gb RAID system, so I'm thinking
> of moving to a HW solution.
Wrong reason. Software RAID is faster. The system CPU probably is a
Pentium II or better, with MMX instructions. It can perform the RAID-5
parity construction at way above the speed of the disks, and probably
if you have an Ultra2 controller or two UltraSCSI controllers, the
SCSI bus bandwidth also exceeds the disk capacity. The resync is
happening at media speed. There's no way a hardware solution could be
faster than that, and low-end devices are very likely slower.
--
Osma Ahvenlampi
