ESXi host logs report repeated logging similar to:
/var/log/vmkernel.log
vmkernel: cpu18:2098658)ScsiDeviceIO: 4633: Cmd(0x45bb8216f900) 0x42, CmdSN 0x641f1 from world 2115282 to dev "naa.###########################" failed H:0x0 D:0x8 P:0x0
VMware vSphere ESXi (all versions)
SCSI command 0x42 is an unmap command.
SCSI code D:0x8 indicates "device busy".
Storage arrays will typically have a limit on the amount of unmap I/O that can be processed.
On some storage arrays, when this is limit is reached, unmap I/O (command 0x42) will be failed with device busy (D:0x8).
If this logging occurs rarely, it may reflect a temporary peak in unmap I/O and can be ignored.
If this logging occurs frequently:
1) review the pattern of unmap I/O from ESXi hosts to the array
2) consider adjusting the level of unmap I/O to avoid overloading the array: Space Reclamation on vSphere VMFS Datastores