![]() |
![]() HP OpenVMS Systemsask the wizard |
![]() |
The Question is: We are running a two-node SCSI cluster consisting of AXP 2100 4/200 under VMS V7.1-2. Both systems are attached to a RA8000 storage system with redundant HSZ70; these devices are seen as $1$DKBxxx. Both systems use PACs (2,3) for their PKA controllers and a host allocation class of 1. Everything works fine until an upgrade to VMS V7.2-1. Then only the boot disk is still named $1$DKB100, all other disks are now seen as $1$DKAxxx! Howzat? We immediately restored VMS V7.1-2 and everything is fine again, but we have to go to V7.2-1. What's wrong ? The Answer is : Nothing. This change is intentional and documented, and parallels previous changes made to OpenVMS to bring all (served) disk device names into the same controller letter. Each device unit number -- regardless of the controller -- must be unique within any non-zero device allocation class. (The FAQ has some additional information on device unit numbers and allocation class settings.)
|