A Drobo Died. Don't Move the Disks.

Written and maintained by the Desert Data Recovery lab team. · Updated October 10, 2026

We already wrote about a controller Rebuild, about Windows Initialize Disk, and about Unraid Format and a TrueNAS pool that will not import. This is a different box. Drobo did not store a standard RAID set that a second controller, Linux, or Disk Management knows how to mount. The drives are a disk pack. The layout is BeyondRAID. The company that built both is not here to take the chassis back.

There is no Drobo left to send it to

StorCentric, the holding company for Drobo, filed Chapter 11 on June 20, 2022. The Northern District of California docket lists Drobo, Inc. in that filing (joint administration with StorCentric, case 5:22-bk-50515; Drobo, Inc. is 5:22-bk-50517). Drobo's own website then said that as of January 27, 2023, Drobo products and support were no longer available, and that support had moved to a self-service model. AppleInsider reported on May 15, 2023 that StorCentric had emailed customers the case converted from Chapter 11 to Chapter 7 liquidation on April 28, 2023. Ars Technica reported the same account the next day.

That is the vendor. It is not a statement that every unit in a closet has failed, and it is not a statement that the files evaporated. It means a chassis that will not boot has no Drobo RMA, no support ticket, and no new firmware. The drives are what is left. What you do with them next is the part that can still be undone, or not.

BeyondRAID is the pack, not a RAID level you can hand to another box

Drobo's 5C Getting Started Guide says that once you insert the drives they function as a unit, a "pack," and that you need the entire pack to access the data. The 5C User Guide says data stored on drives removed from the Drobo cannot be read by other systems. That sentence is the whole problem with a dock and a PC.

BeyondRAID is Drobo's name for how that pack is arranged. Drobo's DroboElite datasheet said BeyondRAID uses the same protection ideas as traditional RAID — mirroring, and striping with parity — applied on a storage virtualization platform, which is how it dropped the limits normally associated with RAID, including the rule that every drive had to match. Drobo's product manual put the distinction directly: BeyondRAID is not a derivative of traditional RAID. A 2009 PC Perspective review of the DroboPro described the drives as a pool of small blocks, with a remap table and other metadata written to every drive, and called the result a proprietary, metadata-driven structure. That review also noted the unit could report a volume to the computer that was larger than the disks actually installed. The map of which block lived where was Drobo's.

That is why a generic enclosure does not "just show the RAID." A PC does not assemble a removed Drobo drive the way it would an mdadm member or a hardware-RAID virtual disk. Drobo's own guide says other systems cannot read it. Published recovery tools do document BeyondRAID support. They are not a reason to attach the original disks to a computer and let the operating system, or the tool, write. The read belongs on images. Windows and macOS are not those tools. A USB dock is not those tools. A second Drobo is not a safe viewer either: reject the pack and the documented button is a reset; accept it and the chassis has write access.

Another Drobo is not a reader

Drobo did document disk-pack migration. The 5C User Guide supports moving an entire pack from a Drobo or Drobo S, or from one 5C to another 5C, both units powered off, firmware already current. The Drobo 8D User Guide lists its own paths — from a 5D3, 5D/5Dt, 5C, Gen 3, or Drobo S — and lists paths it does not support, including Drobo Pro, Gen 1, and Gen 2, and any move between a USB or Thunderbolt unit and a NAS. The 5D3 manual draws the same kind of line. A model pair that is not on the list is not a documented migration.

Where a cross-model move is documented, it is often one-way. The 5C guide says that after a Drobo or Drobo S pack moves to a 5C, you cannot use that pack in the source unit. The 8D guide says the same about a migration onto an 8D. The move also has a precondition the manual is blunt about: do not migrate drives when either Drobo is powered on. That can result in data loss. And the destination is supposed to be on the latest firmware, checked from Drobo Dashboard, before the drives ever move. A dead unit cannot take that update. The company that published the list is not answering the phone to confirm your pair is on it.

If the second unit does not accept the pack, Dashboard's name for the red lights is "Unknown Disk Set." Hardware RAID calls a strange configuration Foreign. Drobo's guides do not. The 5C User Guide's next step for Unknown Disk Set is ERASE (reset). The B810i Getting Started Guide, still hosted on files.drobo.com, says the same and prints the warning in the procedure: it erases all data from all drives and returns the Drobo to defaults. The 5C guide adds that this is for new setups only, where the drives have no data. The pinhole reset beside it is the same format. Clearing the red lights is how the pack gets erased.

If the second unit does accept the pack, it has booted with write access to the only copy. Drobo's data-protection pass is a rewrite. The 5C User Guide says enabling dual disk redundancy will not erase your data — and then tells you data protection is in progress, and not to treat that as idle time. The same guide says that pulling one drive while the unit is powered on gives the Drobo the chance to rearrange your data across whatever remains. Yellow and green bay lights, alternating, are that work. On a healthy migration those steps are the product working. On a pack you pulled out of a dead unit, they are the second chassis rewriting the only copy, with no way back to the source model on a one-way path.

A PC or a USB dock will offer to initialize them

Take the drives out and the computer is one of the "other systems" the 5C User Guide says cannot read them. One disk in a SATA dock is also not the pack. The Getting Started Guide is explicit that access takes the entire set.

What the computer offers instead is a format. Drobo's 5C Getting Started Guide says some Windows and Mac systems will prompt you to format the drives, and the instruction is to click Cancel or Ignore. That note was written for a Drobo appliance sitting on the desk during first setup, when Dashboard — not the operating system — was supposed to format an empty volume. The same dialog on a pack that already has files is the write. Initialize Disk is the Windows version: Disk Management shows a disk it does not trust as Unknown / Not Initialized, and confirming the dialog writes a new partition table. BeyondRAID members are not Windows volumes. Unallocated, in that window, means Windows did not find a table it understands. It does not mean the sectors are empty.

Do not let the dock's own utility "fix" the disk either. Do not run a recovery program against the originals so you can see if a folder appears. A tool that understands BeyondRAID still has to be pointed at copies. The originals stay as they were when the chassis died.

What is actually different from an ordinary RAID

A hardware RAID card, an mdadm set, a Synology volume, and a TrueNAS pool each have a layout a second machine can at least attempt to describe, and each has a button we already tell you not to press. Synology Repair, QNAP Recover, Unraid Format, and zpool import -F are those buttons. They are the wrong clicks on those systems. They are also the wrong instructions here. A Drobo pack is not an mdadm array you import read-only, and it is not a foreign configuration you import on a PERC.

The difference that matters when the chassis dies:

• The redundancy is a setting, not a level printed on the bezel. Drobo's 5C guide protects against one drive failure by default, and against two only when dual disk redundancy is on. Dual disk redundancy needs at least three drives, uses more space, and starts the data-protection pass. Turning it on after the failure is that pass. It is not a repair.
• Drive sizes were allowed to differ. The capacity math is Drobo's, not "N minus one, times the smallest disk." A spare you drop in because it fits the bay is a new drive. The insert-drives warning in the 5C Getting Started Guide says pre-existing data on the drives will be erased, because the drives will be formatted.
• The cache module is not a member of the pack. The Drobo Mini Getting Started Guide, on files.drobo.com, says the mSATA SSD is not part of the disk pack, and that any pre-existing data on it will be erased if you install it. Keep that module with the drives. Do not seat it in another Drobo, and do not leave it in a drawer as if it were a spare.
• Bay order is not what Drobo's migration steps required. Those steps say move all the drives. They do not say the destination reconstructs the volume from your sticky notes. Label the slots anyway, before anything comes out. The label is how a failed bay, a spare someone already added, and the cache SSD stay distinct. It is not a map you can skip and still know what was where when it stopped.

The public case log has no Drobo job. We will not invent one. A Synology RAID 5 and an HPE SmartArray are published on that page. They are different layouts, recovered as those layouts. They are not a prediction about a BeyondRAID pack.

What to do right now

If the bay lights are blinking yellow and green and the drives are quiet, Drobo's shutdown steps say the unit is still protecting data. Wait until those lights are solid green, shut it down from Dashboard if Dashboard is still talking, and then leave it off. Pulling a drive during that pass, or yanking power in the middle of it, is the data-loss case the manual warns about.

If the unit is already dead — it will not stay up, a drive is clicking or grinding, a bay is blinking red, Dashboard is offering Unknown Disk Set, or you already pulled the plug — leave it off. On the 5C, blinking red means Drobo has marked that drive failed. Solid red in the same LED table is a different message, critically low space, and it is still not a reason to pull the drive while the unit is on. Another power cycle is not a diagnostic. See what a clicking drive means.

1. Stop retrying. Do not seat the pack in a second Drobo, same model or a different one. Do not power a Drobo on and then insert the drives. Do not click Reset, ERASE, or the pinhole. Do not enable dual disk redundancy. Do not pull the red drive so the unit can rearrange. Do not put a member in a USB or SATA dock, and do not attach the set to a PC. If Initialize or Format appears, cancel it.
2. Photograph the chassis before the lights change or the drives come out. Model on the label. Which bays are red, yellow, or green. Any Dashboard text still on screen, including Unknown Disk Set. Then cancel whatever dialog is open.
3. Label every disk by slot and serial before it is unplugged. Include a drive the lights already called bad, any disk you already removed, and the mSATA or rear cache SSD if the unit has one. A pile of unlabeled members is how a spare gets mixed into the pack.
4. Keep the whole set. The pack is the unit. A drive that looks failed still belongs with it. The cache module belongs with it. Do not send one disk and a note that says the others were fine.
5. Write down what you remember: the model, whether dual disk redundancy was on, whether anyone already moved the pack, reset it, initialized a disk, or let a data-protection pass run. And whether a backup exists on something that is not this pack.

Pack the drives the way you would any failed drive — how to ship a failed drive safely. If you cannot pull them without guessing which bay is which, the chassis can come in as a chassis. Do not open a drive on a table to see why the light was red.

How a lab handles it

We do not install the pack in a donor Drobo to see if a volume mounts. We do not press Reset to clear Unknown Disk Set. We do not initialize the members, format the volume, or start data protection on the original disks.

Each drive is imaged write-blocked first, including a disk the lights already called bad, as far as that disk will read. The cache module comes along when the unit had one, and it is imaged too, because Drobo's own guide treats it as separate from the pack. Imaging uses the lab tools already named on this site, including the ACE Lab PC-3000 and DeepSpar Disk Imager. Those tools image the disks. They are not a second Drobo. The pack is reconstructed from the copies, and the filesystem Drobo had on the volume is read from that reconstruction.

What imaging cannot do: put back a set a finished Reset or pinhole format already erased, a member a data-protection pass already overwrote, or a disk Initialize plus a new volume already replaced at the start. We will tell you if that is what the copies show. We will not promise the manual's "erases all data" left a full volume sitting underneath a reset that completed.

The public case log has no Drobo. We will not invent one. We don't advertise a success rate, and we can't promise a turnaround from the model name. If the screen in front of you is a PERC or a Smart Array, start with don't force the array online. If it is Initialize Disk on a single drive, start with that dialog. For the work itself, see RAID data recovery.