Workers don't get new IP package from frontier

Problem

During the 3rd run, the workers were throwing exceptions because the frontier stopped answering, e.g.,

{"log":"2017-10-02 13:07:46,583 [main] ERROR [o.h.c.r.RabbitRpcClient       ] - \u003cException while sending query. Returning null.\u003e\n","stream":"stdout","time":"2017-10-02T13:07:46.585655178Z"}
{"log":"java.util.concurrent.TimeoutException: Timeout waiting for task.\n","stream":"stdout","time":"2017-10-02T13:07:46.585709808Z"}
{"log":"\u0009at com.google.common.util.concurrent.AbstractFuture$Sync.get(AbstractFuture.java:269)\n","stream":"stdout","time":"2017-10-02T13:07:46.585717681Z"}
{"log":"\u0009at com.google.common.util.concurrent.AbstractFuture.get(AbstractFuture.java:96)\n","stream":"stdout","time":"2017-10-02T13:07:46.585725113Z"}
{"log":"\u0009at org.hobbit.core.rabbit.RabbitRpcClient.request(RabbitRpcClient.java:156)\n","stream":"stdout","time":"2017-10-02T13:07:46.585732453Z"}
{"log":"\u0009at org.aksw.simba.squirrel.components.WorkerComponent.getNextUris(WorkerComponent.java:83)\n","stream":"stdout","time":"2017-10-02T13:07:46.585739529Z"}
{"log":"\u0009at org.aksw.simba.squirrel.worker.impl.WorkerImpl.run(WorkerImpl.java:80)\n","stream":"stdout","time":"2017-10-02T13:07:46.585746485Z"}
{"log":"\u0009at org.aksw.simba.squirrel.components.WorkerComponent.run(WorkerComponent.java:71)\n","stream":"stdout","time":"2017-10-02T13:07:46.585753514Z"}
{"log":"\u0009at org.hobbit.core.run.ComponentStarter.main(ComponentStarter.java:67)\n","stream":"stdout","time":"2017-10-02T13:07:46.585760435Z"}
{"log":"2017-10-02 13:17:48,586 [main] ERROR [o.h.c.r.RabbitRpcClient       ] - \u003cException while sending query. Returning null.\u003e\n","stream":"stdout","time":"2017-10-02T13:17:48.587168887Z"}
{"log":"java.util.concurrent.TimeoutException: Timeout waiting for task.\n","stream":"stdout","time":"2017-10-02T13:17:48.587222859Z"}
{"log":"\u0009at com.google.common.util.concurrent.AbstractFuture$Sync.get(AbstractFuture.java:269)\n","stream":"stdout","time":"2017-10-02T13:17:48.587230644Z"}
{"log":"\u0009at com.google.common.util.concurrent.AbstractFuture.get(AbstractFuture.java:96)\n","stream":"stdout","time":"2017-10-02T13:17:48.587238083Z"}
{"log":"\u0009at org.hobbit.core.rabbit.RabbitRpcClient.request(RabbitRpcClient.java:156)\n","stream":"stdout","time":"2017-10-02T13:17:48.587245243Z"}
{"log":"\u0009at org.aksw.simba.squirrel.components.WorkerComponent.getNextUris(WorkerComponent.java:83)\n","stream":"stdout","time":"2017-10-02T13:17:48.587252323Z"}
{"log":"\u0009at org.aksw.simba.squirrel.worker.impl.WorkerImpl.run(WorkerImpl.java:80)\n","stream":"stdout","time":"2017-10-02T13:17:48.587259355Z"}
{"log":"\u0009at org.aksw.simba.squirrel.components.WorkerComponent.run(WorkerComponent.java:71)\n","stream":"stdout","time":"2017-10-02T13:17:48.587266352Z"}
{"log":"\u0009at org.hobbit.core.run.ComponentStarter.main(ComponentStarter.java:67)\n","stream":"stdout","time":"2017-10-02T13:17:48.587272977Z"}

This could be caused by an empty queue in the frontier. But this should be answered in a different way. It could be helpful if the frontier regularly logs its status (queue length, number of IPs blocked, etc.).

Trace the timeout from RabbitRpcClient.java:156 through WorkerComponent.java:83 and WorkerImpl.java:80, then inspect how the frontier handles an empty queue. Confirm whether the failure is caused by queue exhaustion and determine what response workers should receive. Done means the frontier exposes useful status such as queue length and blocked IP count and workers no longer wait indefinitely in this case.