Oracle died and was not restarted

During a long run of webindex on EC2. The Oracle died and was not restarted. When I started looking into this I also noticed some workers were not being restarted.

```
10:00:19.194 [ApplicationMasterService] INFO o.a.t.internal.LogOnlyEventHandler - Requested 20 containers for runnable FluoWorker, only got 15 after 19135138 ms.
10:00:49.543 [ApplicationMasterService] INFO o.a.t.internal.LogOnlyEventHandler - Requested 20 containers for runnable FluoWorker, only got 15 after 19165487 ms.
10:01:19.947 [ApplicationMasterService] INFO o.a.t.internal.LogOnlyEventHandler - Requested 20 containers for runnable FluoWorker, only got 15 after 19195891 ms.
10:01:23.019 [ApplicationMasterService] INFO o.a.t.i.a.ApplicationMasterService - Container container_1462212200762_0008_01_000002 completed with COMPLETE:Container released on a *lost* node.
10:01:23.025 [ApplicationMasterService] WARN o.a.t.i.appmaster.RunningContainers - Container container_1462212200762_0008_01_000002 exited abnormally with state COMPLETE, exit code -100.
10:01:23.026 [ApplicationMasterService] INFO o.a.t.i.appmaster.RunningContainers - Re-request the container container_1462212200762_0008_01_000002 for exit code -100.
10:01:23.030 [ApplicationMasterService] INFO o.a.t.i.a.ApplicationMasterService - Container container_1462212200762_0008_01_000020 completed with COMPLETE:Container released on a *lost* node.
10:01:23.030 [ApplicationMasterService] WARN o.a.t.i.appmaster.RunningContainers - Container container_1462212200762_0008_01_000020 exited abnormally with state COMPLETE, exit code -100.
10:01:23.030 [ApplicationMasterService] INFO o.a.t.i.appmaster.RunningContainers - Re-request the container container_1462212200762_0008_01_000020 for exit code -100.
10:01:23.030 [ApplicationMasterService] INFO o.a.t.i.a.ApplicationMasterService - Re-request container for FluoWorker with 1 instances.
10:01:23.031 [ApplicationMasterService] INFO o.a.t.i.a.ApplicationMasterService - Re-request container for FluoOracle with 1 instances.
10:01:49.144 [ApplicationMasterService] INFO o.a.t.internal.LogOnlyEventHandler - Requested 20 containers for runnable FluoWorker, only got 14 after 26113 ms.
10:01:49.145 [ApplicationMasterService] INFO o.a.t.internal.LogOnlyEventHandler - Requested 1 containers for runnable FluoOracle, only got 0 after 26114 ms.
```

Was seeing messages like the following in the YARN RM logs... not sure if related.

```
2016-05-04 04:37:22,167 INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler: Trying to fulfill reservation for application application_1462212200762_0008 on node: worker14:40196
2016-05-04 04:37:22,167 INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler: Skipping scheduling since node worker14:40196 is reserved by application appattempt_1462212200762_0008_000001
```

I attached the Twill AM logs, Yarn RM logs, and Fluo oracle [logs](https://github.com/fluo-io/fluo/files/249045/fluo-657-logs.zip). The application id was `application_1462212200762_0008`. The hadoop version was Apache Hadoop 2.6.3. The Twill version was 0.6.0