Experiments with Drobo 5N during a spot of housekeeping

I needed to swap out a hard drive on one of my Drobo 5N. It all went smoothly as I swapped in a new 6 Terabyte drive. The whole thing only took a few minutes and I left the drive running to hot-swap the WD Red 6TB. No drama.

It reminded me that one of my backup disk packs has 5 x 6 Terabyte drives, but that the available space is only 16 Terabytes. I know, it sounds a lot, but well, there was around another 5.7 Tb of unallocated 'blue' space.

Drobo changed its firmware a couple of years ago to support up to a 64 Terabyte image, but although I use the latest version, my original disk pack predates the 64 TB limit. Time for some reconfiguration and a little experimentation too.

The whole WD Red 30TB pack was a backup of a backup, so I decided to try a couple of experiments before rebuilding. These are kind of 'worse case scenarios' and the kind of things that meander along unanswered on forum posts. I was more interested in the latitude of the Drobo against a human error than in really exercising these varied disasters.

Experiment 1: Reset the Drobo (which should cause it to lose everything) and then re-insert 4 of the 5 disks. Would the Drobo still find the remnants of the pre-existing pack and attempt to rebuild it? Yes, it did, although it estimated around 60 hours for the rebuild. Despite the reset, the revised pack would still be a maximum size 16 TB. Useful to know.

Experiment 2: Shuffle the still existing 5 Pack of disks before re-insertion. The Drobo figured this out too, and attempted to rebuild all 5 of the now jumbled disks. Interesting.

Experiment 3: Erase the 5 disks, one at a time in a USB3 caddy using OSX Disk Utility. Then reinsert four of them into the Drobo and see if it would attempt a rebuild. Yes, even this appeared to work, although I didn't let the 60 hour rebuild run.

Experiment 4: Take 4 of these good disks and another Mac erased WD Red 6TB disk, which had previously been defective in another Drobo. See whether this 'reset' Drobo also decided the disk was defective. Yes, even after the reset I got a red light on the duff disk. The disk works fine in a stand-alone caddy, but in such a mode it isn't subject to the same level of duty as it would be in a server pack. Implies that Drobo maybe marks the disk as well as storing a defective drives table.

Finally, another reset. No drives in the unit. Put in two and wait for the proper re-initailisation to start. New space offered, 64 Terabytes, of which my 30TB pack with redundancy has 21TB to use. That should keep me quiet for a while.