問題文
A cluster has been installed with two management nodes configured as a failover pair. A new operator is asked which of the two is currently doing the work, and answers that it must be the one named as the first of the pair. Why is that answer unsafe to rely on?
選択肢
- Both nodes always do the work concurrently, and neither of the two is idle at any point in normal operation.
- Neither node is doing the work in a failover pair, and all services are provided from the shared storage layer while the two management nodes only monitor each other for availability.
- The two nodes exchange their names during every failover, so after the first handover the node that was named first is now named second and the operator has to re-read the naming before making any change.
- The name of each node is fixed, but which one is currently doing the work is a mode that can change, so the two have to be checked separately.