Why does Scale Logic always use RAID6 or higher and rarely RAID5?

Scale Logic's Genesis RX/UX RAIDs have RAID5, 6, 7, and even N+M where you could have higher than 3 parity drives for data LUNs.  But RAID6 has been the standard for a long time due to its higher reliability in not only the extra drive can be lost but also the integrity of the rebuild process.  RAID6 offers the best protection while still getting decent capacity utiliziation.  Still some of our competitors only use antiquiated RAID5.  Why does Scale Logic standardize on RAID6?  

 

General web information on RAID5 versus RAID6 guides you to think you have disadvantages and advantages with each approach.  But with Scale logic's multi-threaded advanced parrallel RAID calculations there is a very neglectible performance difference in our RAID5 versus RAID 6 so performance is out the window for us.  We have faster RAID6 than our competitors RAID5.  

 

So let's look at the next area people use RAID, they use it for availability. They don't want a drive to die and end up having to take their system or server down or restore data from a backup. To really achieve that, you don't use RAID5 because RAID5 will expose some corner cutting.

Why? Because as drives are getting larger, there's more possibility for URE, unrecoverable read errors. That is a very common failure of a HDD these days and you don't want to discover one or more in the middle of a rebuild.  Scenario: RAID system with 3 drives in RAID5. You get an alarm on your RAID that drive C died. Order new drive or use your spare, swap it out, no problem - right?  In the middle of the rebuild, it dies.  

According to the diagnostics, there was a "bad spot" on drive B.  A unrecoverable read error. But drive B is supposed to be good right?  Wrong it has hit a URE.  The system had a silently failed area on that drive repeatedly, and now that it was rebuilding the data, it couldn't read that spot, and no matter how many times you run the rebuild - it won't rebuild.  The correct solution is to pull the data off and reformat with known good drives and putting the data back on.  Some RAIDs offering RAID5 will knowingly skip this block during the rebuild and continue on like nothing happened...  It isn't banking data thankfully, but what if the area skip contained your Premire Project?  That is a gamble Scale Logic isn't taking.


Scale Logic Genesis doesn't skip a bad block and our RAID6 is faster than most RAID5's.   

Daryl Heinis
2018-10-21 15:33