Comment 2 for bug 1966397

I don't know if the transport protocol for those SCSI volumes is iSCSI or FCP, but those LUNs were most likely used in storage at some point.

If it's iSCSI they are most likely caused by race conditions between concurrent attach/detach operations. This has been greatly improved in cinder, os-brick, and nova and it should no longer be an issue.

Another possibility is that os-brick failed to detach those volumes and Nova ignored these errors leaving dangling devices which still appear in your system.

You should check your logs to see what really happened with the detach of those LUNs that no longer exist on your storage.