Ooops... Nope.
This was a promise Controller. They just called it Raid 1+0 (Is that 10? I thought that was closer to a Raid 5 mirrored)
On Raid 1+0 it was a stripe on two drives that was mirrored to another two drives.
So if you lost a physical drive, it was "Degraded" but the Logical Raid was still alive since you had a "matching stripe" still running. (3 out of 4 drives)
If you lost a second drive later on that happened to be the mirror of the dead drive you were toast. But that's what alerting is all about. But the odds were a little more in your favour. (You would have to lose the twin to the failed to be dead, you COULD theoretically lose and second drive and NOT be dead in the water, I think I've had that on an older Promise controller at least once)
Of course if it's a pure stripe and you lose a drive, well go call CBL right away. But NOBODY would ever put critical data on a pure stripe or span right? Right? (Sorry ... cough.... I've cleaned up consultants who did just that. SQL database for Great plains no less)
Nope I prefer hardware raid over software ANY time (as long as the vendor doesn't do something stupid and funky like intel used to on earlier mirror controllers for IDE)
Software was a wonderful option when Raid Controllers were expensive. But nowadays for an SMB when a pretty much stock $150 Intel motherboard will have at least a mirrored controller, and 500 gig drives are inexpensive; I'll opt for a hardware over software anytime.
