Re: Q: Raid-5 Controller (Hardware aquiration)
> Is it a problem to get a message if a hardware Raid controller
> detects a disk failure?
Yes, it is normally a problem to detect disk failures when using a hw
raid controller. If you are lucky with your choice of raid
controllers, you will find a text file in /proc/ showing the current
status (like /proc/rd/status with DAC960), and the source for a set of
tools to manipulate the status of the raid (adding/removing disk etc).
If you are unlucky, and this is the norm for most raid controllers,
you will not find a way to check the current status, nor have any
tools available to manipulate the raid status.
It would be interesting to gather a list of raid systems, and how they
compare to each other when it comes to detecting status, reporting
problems and manipulating the status from the command line and a GUI
We once tried to contact the vendor of a raid system to get them to
release their tools in a form where we could include them in debian.
This took so long that when the vendor finally had published the
license of their source, it was impossible to buy the raid controller
anywhere. Based on this experience we in the skolelinux project
decided to not recommend using any hw raid where the status info isn't
easily available and the tools haven't got a license making them
compatible with debian main.