According to the Drobolator, a Drobo-S with the Dual Disk Redundancy has a LOWER capacity with 5 x 2TB drives than a Drobo with 4 x 2TB.
This seems counter-intuitive since one would expect the dual redundancy to cost only one more disk than single error correction.
I do hope it is a Drobolator bug, because if that would be true, that would basically destroy any economic advantage the Drobo may have over mirrored disks (which are faster too).
Furthermore it would make the migration of a full Drobo to a Drobo-S impossible if one wants to benefit from Dual Disk Redundancy, which is the big plus of Drobo-S.
Drobolator results :
Drobo w/ 4 x 2TB = 7.3TB = 5.5 data + 1.8 protection
Drobo-S w/ 5 x 2TB w/ Dual Disk Redundancy = 9.03TB = 4.53 data + 4.55 protection
Drobo-S w/ 5 x 2TB wo/ Dual Disk Redundancy = 9.03TB = 7.26 data + 1.83 protection
I think you will find it is right. Before it only had to lose 1 disk but now it has the ability to lose 2 disks and only adding 1 extra bay hasn’t added the extra space you would think it has done.
I still hope it is a Drobolator bug : the usable space deficit vs Drobo is 1TB, definitely not something “minor”.
At that price, very few would use the dual disk redundancy, and in that case, Drobo-S = Drobo, but twice as expensive
It is probably not a bug. The Pro ends up with the same values with 5x2TB. I doubt they would have let such a bug pass them by for this long.
If you think about it a little bit longer I think you will why it is necessary that use that much for protection.
I would also like to add, your argument doesn’t make any sense. “At that price, very few would use the dual disk redundancy.” At that price, pretty much any one is likely to use the redundancy since it was likely the deciding factor between the Drobo and Drobo S.
I think what you meant to say is that few would be willing to pay twice as much for the Drobo S, while gaining only dual disk redundancy. I say only because there seem to be a number of other improvements. Whether they are enough to command the price difference remains to be seen.
Dual-Disk Redundancy (DDR) is more complex than just having two copies of the parity.
Quick example…
A B C D e f
Assume A/B/C/D are data drives, and e/f are parity drives, where e is a mirror of f.
Lose e+f, you are fine, because parity can be recalculated from A/B/C/D.
Lose A+e, you are still fine, because A’s data can be recalculated from B/C/D and parity f.
Lose A+B, now you are screwed. You have C/D, but you can’t recover the data for both A and B, because you’d need at least one of the two to be able to retrieve it.
DDR takes extra management space because to achieve the second level of fault-tolerance, you have to have protect the first level of fault tolerance too.
Well, suneohair, you are right, the Drobolator behaves the same way for Drobo-S or DroboPro.
However, there is something extremely fishy there :
DroboPro w/ 5 x 2TB w/ Dual Disk Redundancy = 9.03TB = 4.53 data + 4.55 protection
DroboPro w/ 6 x 2TB w/ Dual Disk Redundancy = 10.91TB = 7.26 data + 3.65 protection
DroboPro w/ 7 x 2TB w/ Dual Disk Redundancy = 12.73TB = 8.47 data + 4.25 protection
It seems the proprietary “beyond RAID” algorithm of the Drobo works efficiently only with an even number of disk !
Protecting 12.73TB of data wastes less TBs on protection than for 9.03TB !!!
If it is not a new bug in Drobolator (I never checked those configurations before, so I do not know if answers were different before Drobo-S was introduced), it is definitely an ugly (and unadvertised) property of the Drobo “beyond RAID” algorithms.
And it definitely forbids upgrade of a full Drobo to a Drobo-S w/ Dual Disk Redundancy.
So if it not a bug, we have an inconsistency in the DRI offer : the Drobo-S which was supposed to fill a price and performance gap between Drobo and DroboPro should have had 6 slots, not 5, to be capacity-wise compatible with upgrade from Drobo.
Jennifer : could you please confirm those results are correct and there is no bug in the Drobolator ?
Hurrah for Jennifer !
I am much relieved; the Drobo-S vs Drobo capacity deficit is now a bare 80GB, you just have to remember to migrate before you reach the 98% full level… very unlikely for most people, considering the slow-down activated at the 95% threshold.
BTW Jennifer, when will the corrected version of the Drobolator be on-line ? At that time, it is still the buggy one.
All participants to this thread get a Drobo-S exceptional discount proportional to the usable space we recovered for DRI in the Drobo-S : 5.42/4.53 = 20%
Much better than the paltry 6% of the “customer loyalty” program…
