[an error occurred while processing this directive]

HP OpenVMS Systems

ask the wizard
Content starts here

Seeking Details of OpenVMS Galaxy?

» close window

The Question is:

 
I need to find out about the GALAXY. I know that with GALAXY, I can drag and
 drop CPU from one QBB to another QBB. The problem is, do the GALAXY move the
 I/O too? I am afraid the bottleneck still woth the I/O if I move CPU or
 memory? The second thing, if
I have one or more CPU faulty in one QBB, how do I replace it? Do I need to
 shutdown the box or just the QBB? Do the CPU support hot swap?
 


The Answer is :

 
  Please visit the OpenVMS Galaxy portion of the OpenVMS website, and
  please contact Cerner for information on Cerner products and Cerner
  recommendations.
 
  Various types of I/O controllers can have their interrupts targeted
  to specific processors, this is known as Fast Path.  So yes, you can
  potentially (re)target your I/O across available processors.  (You
  will want to consider the configuration from the individual partitions,
  and view the transfer of a processor from one partition to another as
  simply issuing a STOP/CPU in one partition, and then using START/CPU
  in another.  There is far more involved here of course.)
 
  The specifics of swapping and power-down requirements depend directly
  on the particular processor platform involved.  OpenVMS Galaxy is also
  not a hardware fault-tolerant environment; a processor hardware fault
  can potentially cause the failures of one or more software instances.
  (The console will then configure around a hard failure of a processor,
  and will permit the system(s) to reboot with the remaining processor(s).)
 
  The specifics of the Fast Path I/O also depends on the OpenVMS version
  and the specific controllers involved -- specific high-volume controllers
  have support for Fast Path on specific OpenVMS versions.
 
  The OpenVMS Galaxy configuration utility (GCU) is aware of Fast Path,
  and will not permit a processor that is a Fast Path target to move.
  Further, the GCU is also cogniscent of what components need to be
  powered down within a particular platform, should component hardware
  need to be swapped.  (In older hardware systems running OpenVMS Galaxy,
  the power-down domain is always the entire system.)
 

answer written or last revised on ( 24-JUL-2002 )

» close window