Are your LUNs mapped out from a DS4300 array but not seen by host

Some logical devices from a DS4300 DiskArray (FAStT) were mapped out to an Xseries 336 server that was using qlogic qla2xxx drivers (v8.0.2) for the HBA cards, and RDAC (aka MPP) drivers (v09.01.B5.04) for multi-path failover.

Using the hot_add feature of the RDAC system some new scsi devices corresponding to the mapped out LUN's on the array were created on the server. After a reboot all the FAStT logical devices disappeared, and when re-running the "hot_add" command it failed to see any of the Luns. Using the SMclient on the FAStT all the LUNs were present and mapped out to the server.

The following errors were recorded in /var/log/messages when attempting to re-add the devices using hotadd:

Jan 24 17:22:36 host1 kernel: qla2300 0000:08:02.0: Scheduling rescan for new luns...

Jan 24 17:22:41 host1 kernel: qla2300 0000:06:01.0: Scheduling rescan for new luns...

Jan 24 17:22:53 host1 kernel: 62 [RAIDarray.mpp]Device 0xf40a6400 is not an array

Jan 24 17:22:53 host1 kernel: 62 [RAIDarray.mpp]Device 0xf40a6400 is not an array

Jan 24 17:22:53 host1 kernel: scsi: Unexpected response from host 4 channel 0 id 0 lun 8 while scanning, scan aborted

The reason for this problem was on the DS4300 array. The LUNs that were mapped out did not have consecutive LUN IDs starting from zero. If the LUN IDs start from a number higher than zero or have gaps in the sequence i.e. 0,1,2,3,6,7,8 then the host will fail to recognize the LUNs.

To fix reorder the LUN ID numbering on the array