Disk size bouncing between normal and zero due to returned SCSI success for failed READ CAPACITY command

Issue

• Disk size is repeatedly changing and bouncing between normal/expected size and 0 size.

  kernel - - - sd 2:0:1:1: [sdbf] Sector size 0 reported, assuming 512.
  kernel - - - sd 2:0:1:1: [sdbf] 0-byte physical blocks
  kernel - - - sdbf: detected capacity change from 15986853937152 to 512
  :
  kernel - - - sdbf: detected capacity change from 512 to 15986853937152 
  

• This results in IO being rejected as it references sectors beyond the end of disk resulting in data loss.