Disk Eject Errors

Note: in section 6 check the sleep locations for Sequoia and Tahoe. I may also add a terminal command that forces system not to sleep, but need to write that.

What is a disk eject error?

When a SoftRAID volume is mounted and a disk disappears or ejects unexpectedly, SoftRAID will notify you of this error.

Common Causes and Solutions

1. Enclosure disconnected or powered off while volume was mounted

Cause: User inadvertently disconnected or powered off an enclosure before properly unmounting the volume.

Solution: Always unmount volumes completely before powering off or disconnecting enclosures. Wait up to 2 minutes after unmounting to allow drives to flush their onboard cache to disk.

Shape

2. Volume unmounted but macOS did not fully release it

Cause: Applications (especially Time Machine) or macOS processes (like Spotlight) did not fully release the volume after unmounting. The drives may still be flushing cache, or an application may be holding resources open.

Note: Time Machine can keep resources open on volumes even after unmounting. This is a macOS bug that has mostly been fixed in Sonoma but can still occur. Spotlight indexing issues can also cause this.

Solutions:

For Time Machine volumes:

• In Time Machine preferences, uncheck “Back up Automatically” and test if this prevents the problem

For non-Time Machine volumes:

• Go to System Settings → Spotlight
• Temporarily exclude this volume from Spotlight indexing
• Restart your Mac
• Unmount the volume and wait 2 minutes before powering off
• If this solves the problem, re-enable Spotlight (it will rebuild the index)

3. Thunderbolt disconnection event (most complex issue)

Cause: An electrical noise event causes the Thunderbolt connection to momentarily drop. Thunderbolt is hot-swappable, so when the connection drops, all disks instantly power down until the connection re-establishes. The volume typically remounts automatically, but SoftRAID warns you about the ejects.

Technical explanation: Every Thunderbolt connection point has two controller chips (one on the cable end, one on the device) that communicate continuously. If either chip loses signal, crashes, or resets, it auto-disconnects from the Thunderbolt bus and powers down the disks.

Most users never see this problem. It is a design limitation of Thunderbolt and can be difficult to troubleshoot. The more Thunderbolt devices connected to your computer, the more likely this issue will occur. Even a Thunderbolt monitor can trigger disk ejects in some cases.

Solutions (try in order):

• Check all connections - Ensure all Thunderbolt cables are firmly connected
• Check for cable movement - Ensure pets or environmental factors aren’t moving cables
• Test different ports - Move cables to different Thunderbolt ports one at a time to identify which part of the Thunderbolt bus is causing the issue
• Try a different Thunderbolt cable - If you have a spare, swap it to rule out cable failure
• Bypass Thunderbolt docks - If using a Thunderbolt dock, try connecting the enclosure directly to your Mac for a few days
• Disable sleep settings: macOS Monterey:
• Set display sleep to “Never”
• Uncheck “Put hard disks to sleep when possible”
• macOS Ventura/Sonoma:
• System Settings → Lock Screen → Set “Turn display off when on power adapter” to “Never”
• System Settings → Displays → Advanced → Check “Prevent automatic sleeping on power adapter when the display is off”
• System Settings → Energy Saver → Uncheck “Put hard disks to sleep when possible”

Note: Disk ejects like this are very rare with USB. If a USB enclosure keeps ejecting, it is a cable issue or failing enclosure.

4. Single disk ejects from a multi-drive volume

Cause: If one disk ejects while others in the same enclosure remain connected, this is most likely a disk error, early sign of failure, or an enclosure slot issue.

This section applies when a disk ejects then reappears shortly afterward.

For HDDs (rotating media):

When only one disk in a multi-drive enclosure (like a ThunderBay) ejects, this is a hardware issue with either the drive or that specific enclosure slot. (It’s possible for 2 adjacent disks to eject if they share the same internal bus.)

Troubleshooting steps:

Assuming SoftRAID shows no SMART predicted failure warnings:

• Identify the ejecting disk:
• Open the SoftRAID log to find the SoftRAID ID of the ejecting disk
• If you cannot identify it, give each disk a unique SoftRAID label – the log will show the label next time it ejects
• Swap the disk to a different slot:
• Important: Avoid the slot directly adjacent to the current location
• Enclosures typically have drive pairs that share the same internal PCIe bus
• Example: In a 4-drive enclosure, if the ejecting disk is in slot C or D, swap it with a disk in slot A or B (not the adjacent slot)
• This isolates whether the problem is the disk or the enclosure slot
• Interpret results:
• If the same disk ejects in the new location → Replace the disk
• If a different disk ejects in the original slot → The enclosure is failing

For SSDs/NVMe (flash media):

Flash media can sometimes “hang” due to early failure modes or overheating. However, if this happens, the drive typically does not immediately reset and reappear in SoftRAID.

Solution: If a flash drive in a stable enclosure ejects and reappears, replace the drive.