scsi-stop in RAID1

[s1za]

I'm using two 1tb discs in RAID1 (using mdadm) and mounted as md0

$ cat /proc/mdstat
Personalities : [linear] [raid0] [raid1] [raid5] [multipath]
read_ahead 1024 sectors
md0 : active raid1 scsi/host0/bus0/target0/lun0/part1[0] scsi/host1/bus0/target0/lun0/part1[1]
      976759936 blocks [2/2] [UU]

if I use the scsi-stop script (one for each disc) the discs spindown correctly but when the discs are needed again only one disc resume, breaking the RAID sync. Is there a way to stop this from happen or another script to use raid md0?

scsi-stop.sh /dev/scsi/host0/bus0/target0/lun0/disc
scsi-stop.sh /dev/scsi/host1/bus0/target0/lun0/disc

thank's in advance.

[s1za]

Can someone give me an opinion...
Thank's

[Serpent]

In my opinion you are asking too much from an software RAID!!!
I'm currently using SW and HW RAIDs (on Solaris), but only HW ones have power saving capabilities.

[s1za]

The problem isn't the RAID software, this works flawlessly.
The problem is that 'scsi-stop' expects a scsi device (/dev/scsi/host0/bus0/target0/lun0/disc) and doesn't accept the 'md0' device (which has the RAID).

'scsi-stop' puts both discs to sleep (again flawlessly according to log files) the problem is to wake-up and only one disc resumes. Is there a turn around to this?

Thank's