SoftRAID: macOS Hang, Kernel Panic (12 Disks)

Created on: April 22, 2022
Last updated: April 12, 2024

Recently the SoftRAID team identified and was able to reproduce a bug on macOS (so far confirmed in 10.15.x) where if a customer has 12 or more disks connected to the host — the number and type of volumes doesn't appear to matter — the system will hang, and in many cases a crash (i.e. kernel panic) will follow. For example, the SR team was able to reproduce the crash by creating a 16 disk SoftRAID or AppleRAID stripe (RAID 0) volume, and then running a continuous AJA System Test on the volume. A kernel panic usually results in as little as 2 hours of disk use under these conditions. Additional details are listed below. 

• This bug is seen only with high amounts of disk activity, on systems connected to a total of 12 or more SATA disks.
  • The disks do not have to be part of the same volume / a single volume.
  • Most commonly seen when a customer has 3 or more ThunderBays in use.
• This bug affects both AppleRAID and SoftRAID volumes.
  • This bug does not affect solutions that are based on NVMe M.2 drives (e.g. ThunderBlade, Express 4M2, or Accelsior 4M2).
• When the user launches SoftRAID immediately after the hang, one of the disk tiles will freeze with the progress bar being displayed in the "getting disk info" step.
• This bug has been reported to Apple; we do not have an ETA but are hoping for a fix in 10.15.5.