This message was deleted.

I don’t see the following on

ThreadingTaskRunner.java

final File logFile = new File(taskDir, "log");
final File reportsFile = new File(attemptDir, "report.json");

In indexer mode the logs are routed to the

indexer

log itself, not the individual task VMs (which is where the UI is attempting to fetch from I think). We send out

indexer

logs to stdout and collect to Kibana, and that contains all logs for the entire node.

OOOOOOO that’s a big change.

Yeah it took me a bit to locate. Since the indexers are all just running in a threadpool inside the single VM all logs go to one place.