Occasionally you need may boot a server without first powering on a tape shelf, or even just an external tape drive. Traditionally most people will resolve this issue by simply rebooting the server in question.
This is a perfectly fine solution for a backup server or even a standalone server that has very little use. Now what happens when you have told everyone the server is fine and you have over 100 users logged in? Do you schedule another outage? or just make the drive work?
On Redhat just issue the following command to make the system rescan the SCSI bus
echo “- – -” > /sys/class/scsi_host/host0/scan
Running dmesg will display the tape shelf / drive / device and will be usable.