I have a bladecenter configuation which I am building out for a customer to vitulize their entire IT infrastrucute. I setup the SAN to provide a 5GB LUN (0) to each specific blade (total of three LUNs) plus one VMFS0 1.6TB volume (15). Each server has a daughter card which is a Q-Logic dual port FC HBA. I have recorded the WWN and setup the LUNS so that the two WWN for each blade is associated with their corresponding boot LUN (0) and then all (6) WWN for the three blades see the VMFS0 Volume (15). I have gone into the BIOS of each blade and setup the boot target for each of the first ports on the Q-Logic to scan for and peg LUN 0 (5GB with correct WWN corresponding to the host). Systems are all configured to boot from the PCI adapter as primary (onboard SAS disabled) and to get the boot CD to run I press F12 to select boot option.
Their are additional VMFS volumes but at this time they are not in the storage groups or consideration at this time.
ISSUE:
Blade1 boots up and sees both drives but the enumeration is /dev/sda (1.6TB) and /dev/sdb (5GB) --->Wrong
Blade2 boots up and sees both drives with /dev/sda (5GB) and /dev/sdb (1.6TB) ---> Correct
Blade3 boots up and sees both drives but the enumeration is /dev/sda (1.6TB) and /dev/sdb (5GB) ---> Wrong
What I have done:
1) Checked BIOS settings to try to find inconsistancey(it was late so I will review this once more this afternoon)
2) Check /proc/SCSI to see if their are differances in the scans that it list which their are .... but I do not know HOW to effect this.
3) Check Q-Logic Settings to ensure their are not any settings differant (but I will also be rechecking this)
I do not understand why the systems are inconsistant and though the blades will load out the systems to boot and such, they are not consistant which in the case of a cluster/ farm, their is concern on WHY this is and how to make it at a minimum consistant (if not correct). I am sure it is somthing stupid, but hopefully someone has experiance in similar acts of stupidity.
Thanks,