Software RAID will work a lot better then any kind of rsync setup. It is really a pretty simple setup, they create a virtual device for the disk /dev/mdx usually. There is linking done so that when you write to the md device it actually writes to all disks in the device (for mirroring... it does other things for more complex raid modes). The code is actually pretty straight forward and easy to understand. I've run linux software raid for about 2 years now and I haven't had a problem yet. It also uses the usual kupdated routine to write to disk so if you do have a disk die you can rest assured that the 2nd disk is going to be an exact replica of the first.
What do you mean by "scan"? Does it actually check to see if the file system is clean and run fsck? Does it actually mount the disk?
If the disk is in the system and plugged in the kernel will recognize it (unless you disable the IDE controller in the bios). It will show the device, and the partition layout.
Something like this:
Code:
hde: attached ide-disk driver.
hde: host protected area => 1
hde: 234441648 sectors (120034 MB) w/8192KiB Cache, CHS=14593/255/63, UDMA(100)
Partition check:
/dev/ide/host2/bus0/target0/lun0: p1 p2 p3 p4
If you have the file system in /etc/fstab linked to a mount point then it is going to get mounted at book. If the file system dirty bit is set then it is likely your init scripts will automatically run fsck on the file system.
You could put the noauto flag in your fstab, which would mean it wouldn't mount the disk automatically on boot. If you have intelligent init scripts it should also not bother checking the disk.