Intolerant RAID?

When I read “reviews” from sources such as CNET that are so blatently incomplete, false and incompetent as this, I’ve got to speak up:

Direct-attached external storage just doesn’t get any better than Maxtor’s OneTouch III Turbo Edition.

The evidence is pretty clear — CNET gives this device a glowing review, yet the 19 user reviews on the site completely trash it.

My own experience is similar to these 19 users. We picked up one of these devices as an easy external RAID solution to provide redundant backup on our server. I have to admit, the form factor is great, but that’s where the greatness ends and poor design and implementation pick up.

We had two problems with the device, both of which are show-stoppers in my mind. First, whenever the server restarts the drive fails to publish its availability — in other words, if the server restarts, you have to physically turn off the drive or it won’t reconnect to the server. But this deficiency pales in comparison to the device’s uselessness as a fault tolerant backup device. RAID is suppose to be tolerant of system failures. We used the device in a mirrored configuration so that if one drive failed the system continues to operate. In fact, one drive did fail — and subsequently the device completely failed to initialize.

That’s right, our Redundant Array of Independent Disks, designed for fault tolerance and recovery, failed to recover when one of the mirrored disks failed. I might as well point out that the drive failure occurred within nine months of purchase, which is a pretty poor track record in and of itself.

This is the second time we’ve been bitten by poor Maxtor technology. My advice: Don’t buy Maxtor. At a price tag of $800 this device should not be a low-end component — but Maxtor has delivered from the bottom of the bucket.

On a broader note, shame on CNET for not bothering to publish a review that at least tests out the hardware to ascertain its value. CNET reviews are little more than brochure-ware for their advertisers. What happened to the days of consumer product testing? Would it be too much to ask that systems be put through trials before earning a CNET “Editor’s Choice” award? CNET, if you’re listening, take a note from Consumer Reports and actually find out how effective these products are before telling your readers to rush out and buy one. It would have been easy enough to yank a drive cable and see if Maxtor’s fault tolerant device was, in fact, fault tolerant.