Currently, BookKeeper uses hostname or default IP to calculate index of current node vs all Rubix nodes. However, this doesn't always work in more sophisticated deployments, where Presto nodes use neither to announce themselves to coordinator. Therefore ClusterManager should provide getCurrentNode or getCurrentNodeIndex that would work in such cases.
Currently, BookKeeper uses hostname or default IP to calculate index of current node vs all Rubix nodes. However, this doesn't always work in more sophisticated deployments, where Presto nodes use neither to announce themselves to coordinator. Therefore
ClusterManager
should providegetCurrentNode
orgetCurrentNodeIndex
that would work in such cases.