The orchestrator should be able to tell the batch system scheduler
when a node becomes available/unavailable.
I think this requires a few modifications to the `Orchestraror` code
and the `BatchSystem` interface:
1. Modify the `BatchSystem` interface so that it has two additional
methods `enable(nodename)` and `disable(nodename)`, which by
default do nothing.
2. Implement the two methods in the `batchsys.GridEngine` class: they
should run the appropriate `qmod -e` and `qmod -d` SGE commands.
3. Modify the `Orchestrator` code so that
`self.batchsys.enable(nodename)` is called when a VM signals it's
READY, and `batchsys.disable(nodename)` is called when a VM is
about to be stopped (goes into STOPPING state).
Original issue reported on code.google.com by riccardo.murri@gmail.com on 26 Apr 2012 at 9:53
Original issue reported on code.google.com by
riccardo.murri@gmail.com
on 26 Apr 2012 at 9:53