Drobo

Used space seems wrong

Preping to add another Drobo (5N2 this time) and I used Drobo Dashboard to power down a 5N. I moved the 5N to a new location and powered on. Instead of seeing 10% unused space I now get 60%. I’ve used compare programs to compare backups and nothing seems lost. I’ve spot checked files and they seem to all be OK.

Has anybody ever seen anything like this. The Drobo has 2 12TB and 3 8TB drives and was reporting 10% space unused. Is this just a problem with reporting space or do you think I just haven’t found the missing 16TBs yet?

Very strange indeed…

That is strange. No other indications of failure?

No other problems. Drobo Dashboard shows the correct drives and capacities and comes up with 32.55 for 12+12+8+8+8 drives. The drobo shows 4 lights indicating about 40% and it shows the 4 lights in Dashboard. I’m assuming that the disk space and free space shown are reported by the 5N also.

A year or so ago I did a factory reset to gain the 64 TB partitioning of the drives. Assuming the drive prices would fall for large drives and it would be cheaper to add space rather than buy another NAS (I was wrong so I went with a 5N2). So it has the latest firmware and partition structure for a 5N.

I have been using GoodSync to copy work from my computer to the 5N. I have a 5D at my desk so I can do direct comparisons. The comparison between the main directories show no differences. I use TreeSizeFree to give me a look at the the size, files, and directories. And it says that all of the files are there. I’ve done some spot checking and they all seem to be intact with all of the real data. I plan on doing byte by byte compares (it’s only time) to verify I haven’t lost anything.

The oly thing I can think of is that I crossed some 32 TB threshold (I was down to 10% free space) and then powered down to move the 5N then When I powered on, this problem came up. In the event no one answers or has any ideas I was going to move some of the data to the new 5N2 and then reboot the 5N to see if I can UN-cross the line.

My other thought was to get a report from the drobo itself as to the size of each share and compare with the TreeSize program. I don’t know how to get a report like this so if you have any ideas…?

My other thought is that maybe the size of the largest share which is about 29TB may have grown to large?

If you can think of any ideas I’d appreciate it. Other wise I’ll have to do a hardware reset and try putting everything back.

I have moved about 3.8 TB from the 5N to the new 5N2. That left 25.3 out of 32.5 TB on the 5N and I rebooted. Magically the 16TB of free space dissappeared. SO the FreeSpace calc and the blue LEDs again reflect the amount of space on the 5N. I will be doing more checking, but at this point I’d have to conclude that as you approach 32TB the space calculations have a bug in them.

As I moved the 3.8 TB from the 5N the free space continued to climb. After the reboot the “lost” 16 TB went back into the UsedSpace. 1-This indicates that the space is calculated based on the initial state of the disks. 2-The 32 TB and 16 TB numbers (doing the math in my head) seems to indicate an error on the boundry of a 47 to 48 bit number. A roll over in the high order bit of 3 16 bit words?

I don’t know for sure, but for now I’m going to treat anything above 85% used space as a thing to be avoided.