When a LUN removed from the storage side you need to scan for changes on your host. If you won't remove the dead lun path system shows them in NO_HW state. This means the operating system no longer probe for LUNs on array controller so the LUNs behind the controller go to an unhealthy state or NO_HW state. When you get a LUN from the storage side which LUNID is the same as the old disk which removed before, your system is not able to check or detect this new LUN. You have to remove the old NO_HW state path then scan again.
Step 1: List all NO_HW device.
#ioscan -fnC disk| grep -v grep | grep NO_HW | awk '{ print $3 }'
Check if there exists a process that running on this device. If not then use "rmsf " command to remove the NO_HW device.
Step 2: Remove NO_HW device
#fuser /dev/rdisk/diskname #rmsf -k -H <diskpathfromioscanfncdisk>
A simple script that removes NO_HW device. Be careful before running this script. It 'll delete all NO_HW devices.
#for i in `ioscan -fnC disk| grep -v grep | grep NO_HW | awk ‘{ print $3 }’`
do
rmsf -k -H $i
done
