I have what seems like a simple test, but have a hard time finding a solution.
I want to check a Switch stack that has a few simple rules.
If stack is up stack IP is pingable.
When stack is down (due to one switch failure) one of the independant IPs is pingable.
So I'd like a test to ping the stack and ping the nodes when the stack is down.
I've established host dependencies that do set the dependent hosts to pending when stack is up and test the stack nodes when cluster is down, this is ok but !
Once the test was run and the state was changed to Down, even if the stack is repaired and its ip restored to up state, the nodes are no longuer in a pending state but in a down state.
How could I have the state reset to pending or similar ?
I've also run tests with dependent services and volatile, but no success until now.