gmirror or ata problem

[Søren Schmidt]

Eric Anderson wrote:
>
> Is there any reason the sleep and idle pieces of ataidle could not be 
> added to atacontrol?
>
Yes, it is a dirty hack solution until the idle/sleep thing is 
implemented correctly in the ATA driver framework as just sending the 
command(s) to the devices doesn't quite cut it without tinkering with 
the timeouts which leaves fx server users with suboptimal delays on 
failures.

Anyhow I now have a full time disk server here that could save quite 
some power when doing this, so it just might level up on my TODO list :)

-Søren