OWC Gemini Thunderbolt RAID not mounting drives after restart — requires full power drain

Thanks for sharing that — unfortunately, that mirrors my experience almost exactly.

For what it’s worth, I ended up returning my Gemini and replacing it with a Mediasonic 4-bay enclosure. It mounted immediately, survives reboots without needing power removal, and has been flawless so far — at roughly half the price.

What still puzzles me is how OWC continues to sell the Gemini given the mounting and power-state issues multiple users are seeing. Functionally, it behaves like a USB 3 enclosure bridged to Thunderbolt, with none of the reliability or advantages you’d expect from a true Thunderbolt storage device.

After switching enclosures, the problems disappeared entirely, which strongly suggests the Gemini itself is the weak link.

I’m still going back and forth with OWC support trying to get a straight answer on whether this behavior is expected for the Gemini. They consistently avoid answering that question directly. Rather than recommending a refund, they’ve suggested I send the unit to them for testing instead. At this point, it feels like they know this is simply how the device behaves and are unwilling to say so plainly.

My new drive:

That’s exactly what I was seeing.

In my case, I didn’t have to shut down the Mac — I just unplugged the Gemini’s power completely and let it sit before reconnecting it.

The tricky part is the timing. It wasn’t instant. The first time I unplugged it, I waited maybe 20 minutes and it still didn’t show up when I plugged it back in. I unplugged it again and waited longer — closer to 30–60 minutes — and then the drives came back.

So unfortunately I don’t have an exact “wait X minutes” answer. It seems like the internal controller needs a full discharge/reset before it will reinitialize properly.

Once it came back, my data was intact — it was just the enclosure failing to bring the drives online after restart.

Let me know how long you end up needing to wait and whether it returns. The more data points we gather, the clearer the pattern becomes.

Thanks for sharing this — and I’m really glad your data turned out to be intact. That must have been a stressful moment.

What you described sounds very close to what I’ve been seeing: the enclosure itself remains functional, but the internal drives fail to initialize after certain events (restart, power state change, etc.). In my case, the only reliable fix was a full power drain of the Gemini before reconnecting.

The part that really stands out in your experience is that both drives appeared dead — even when tested externally — and then later spun up normally. That suggests the drives themselves weren’t actually failing, but were stuck in some kind of initialization state.

For comparison, I eventually tested the same drives in a Mediasonic 4-bay RAID enclosure, and I haven’t been able to reproduce the issue there. That doesn’t necessarily prove where the fault lies, but it did make me more confident that the drives themselves were not the root cause.

Out of curiosity:

• Are you on Apple Silicon or Intel?
• Did this happen after a reboot or an unexpected shutdown?
• Were you using Thunderbolt only, or also USB?

I agree — on paper the Gemini looks ideal. That’s why I bought it. But once you see something like this, it’s hard to fully trust it for long-term storage. I did reach out to OWC about the behavior, but so far I haven’t received a technical explanation for what might be causing it.

Thanks again for posting — this helps more than you probably realize.

Thanks for laying all that out all of this is painting a clear picture.

The unexpected shutdown detail is interesting, but the fact that your other devices (Drobo, Mac Mini, etc.) were unaffected makes it harder to attribute this to a general power event. And the behavior you described — green LEDs, slight internal noise, but no actual drive spin-up — is very similar to what I saw when the controller appeared to be in a stuck state.

What really stands out is that the drives themselves continue to work perfectly in an external SATA dock. That strongly suggests the disks and RAID metadata were intact, and that the enclosure simply failed to properly reinitialize the storage layer.

I completely understand not wanting to keep tearing it down and experimenting with wait times. At some point, reliability matters more than diagnosing the edge case.

I also get the appeal of the all-in-one Thunderbolt + USB + HDMI hub design — that’s what drew me in as well. I actually still have the older Gemini (the DisplayPort version), and it’s been solid for me, though it tops out around 550 MB/s, which limits its usefulness for RAID 0. The added HDMI output and higher throughput of the newer model made it seem like the perfect upgrade on paper. It’s a clean, elegant solution. But once you run into reliability concerns, it becomes much harder to fully trust it for long-term storage.

It is maddening that OWC does not disclose this limitation.

Glad you were able to recover everything before returning it. That’s the important part.

That’s really interesting — especially the 48-hour power drain part. That’s far longer than what I needed, but it reinforces the idea that the controller can get stuck in a state that only clears after a full discharge.

It’s also notable that the Terramaster mounted the RAID0 volume immediately. If both enclosures are using the same ASMedia hardware RAID controller, that suggests the array metadata itself wasn’t the issue — it was likely the Gemini failing to properly reinitialize the controller after certain events.

The fact that it failed again immediately after an eject is also telling. That points more toward a controller reset/firmware behavior than anything happening at the macOS level.

Thanks for sharing that — this is exactly the kind of comparison that helps narrow it down.

Thanks for taking the time to follow up with all of that detail — that’s extremely helpful.

The part that stands out most to me is the delayed recovery after an extended power-off period. That lines up with what I saw as well: the enclosure would appear completely unresponsive until it had been disconnected long enough for whatever internal state it was in to fully clear.

It’s also interesting that your RAID 0 array mounted immediately in the Terramaster enclosure. That suggests the array metadata itself was intact and that the drives weren’t actually failing — which makes the behavior look more like a controller initialization issue than a disk problem.

For what it’s worth, I’ve also used other OWC Thunderbay units without incident. That’s part of why this was so surprising to me. On paper, the Gemini looks like a solid solution.

At this point, I’m mostly interested in understanding whether this is:

• a firmware edge case,
• something specific to certain macOS / Apple Silicon power states,
• or a controller reset behavior that isn’t handling restarts properly.

I’m glad you were able to recover everything before returning it. That must have been a relief.

Thanks for the follow-up. At 40+ hours fully disconnected with the drives removed, that really does rule out a simple power-state hang. At that point I’d probably make the same call you did.

Splitting the duties between a dedicated enclosure and a separate Thunderbolt dock actually makes a lot of sense. It reduces the number of variables tied to one single device. And you’re right — Thunderbolt’s daisy-chaining is one of its big advantages. It does feel a bit like the old FireWire days in that respect.

On paper, the Gemini concept is elegant. In practice, modular setups can sometimes be more predictable. I hope the TerraMaster + Thunderbolt Go Dock combo works out smoothly for you.

Appreciate you sharing the outcome — it helps others reading this thread.