SLAC CPE Software Engineering Group
Stanford Linear Accelerator Center
System Admin

How to failover From



SLAC Detailed
SLAC Computing
Software Home
Software Detailed

Last Modified: February 19, 2010

Booting up MCCDEV's Hardware as MCC:

  1. All cluster members must be down
    1. Type: show cluster
    2. MCC gets shutdown last, so shutdown all nodes before MCC
  3. Switch the SLCNET PCIL fibers between MCC and MCCDEV hardware
    1. The fibre cable pair that connects the PCIL to the SLCNET Gateway is bright yellow with a blue connector.  The PCIL is in the second-from-the-top-most PCI slot, and the fibre connects to it from the back of the Alpha(s).  Simply pull (gently) the fibre out of its connector on MCC, gently reroute it to the back of MCCDEV, unplug MCCDEV's PCIL fibre, and plug MCC's PCIL fibre into MCCDEV's PCIL.
  1. Boot the MCCDEV hardware from the MCC root directory on the system disk, and start the production control system.
    1. At the ">>>" prompt:  boot   -flags  4,0

Subsequently you could reboot any other cluster members (except failed MCC hardware) normally.  For testing the MCC hardware, you could reboot it from the MCCDEV root directory on the system disk ( boot -flags 5,0 ). After hardware failures have been fixed, you could choose to leave the roles of MCC and MCCDEV permanently swapped, rather than going thru a similar FAILBACK procedure.



boot_dev    dua0.

boot_osflags     4,0

boot_dev    dua0.

boot_osflags     5,0




[SLAC CPE Software Engineering Group][ SLAC Home Page]