When you replace a service controller at the
same time that you replace the disk drive, you might not be able to
perform a node rescue because the nonvolatile memory
in the "new" service controller does not contain the operating
system software required to do so.
Also, if you must replace the hard disk and
the service controller at the same time, you cannot boot the node
to perform node rescue. If you find that you cannot run node rescue,
then use the following procedure to perform the rescue:
- Swap the service controller with a service controller from a working
node. The results are the following:
- The "new" service controller that is swapped into the working
node has its nonvolatile memory updated when the node is booted from
the hard disk.
- The service controller that is swapped into the failed node from
the working node contains the operating system that is required to
perform node rescue on that failed node.
- When all updates are complete, swap the service controllers again,
returning them to their original nodes.