with RAID 5 you have drive redundancy, but with RAID 6 it distributes the data to rebuild your array across the volume in such a way where you can lose two whole hard drives (but not three) and still rebuild.
With larger drives you have higher likelihood that you’d have your failed drive, but then the rebuild data from the other drives might fall on a bad sector, meaning you’d lose all your data anyway, that’s why everyone wants RAID 6 nowadays. Yes, it take a long time (like hours, maybe many) to rebuild a RAID 6 after swapping drives, it all happens behind the scenes and you can use your RAID again as soon as you insert the new drive back in the box again, it’ll just be a bit slower to read/write while it’s magically rebuilding for you. 15K SAS drives are awesome and fast, but you can add lots of commodity SATA drives for a LOT cheaper and decent I/O because you make up for slower drives with more spindles turning to give better read/write performance. That’s what a lot of the bulk storage guys are doing these days. YMMV. On 12/6/17 12:02, Adam Moffett wrote: > Well...I'm using 15K SAS in RAID10. Only 600GB because enterprise drives > are $$. > > Out of curiosity what is it that I would need good luck for? > ?Reliability problems? > > The largest RAID6 I have currently is 12x 4TB with a hot spare. It takes forever to rebuild, but it hasn't ever failed outright. The risk is that another drive barfs while rebuilding.