Hi there guys,
I’m starting a middleManager on four nodes that have the exact same configuration file (except for the druid.host entry that has their respective hostnames) - in three of them it goes up but in one of them (cluster-a, same place where the overload and coordinator are), it fails with a message: “druid.indexer.logs.directory - may not be null”).
The problem is that this entry is not null in the settings, it points to local directory that exists on all nodes (using local storage, not remote). Any ideas?
For reference: http://pastebin.com/2rSwSTkE