re-installing the framework

[sadovnikov]

Somewhat related to #83, but it's a different problem - now the Scheduler is not available, when Executors are stopped. The user is trying to reinstall the framework with different code or parameters

The situation:

• scheduler has been stopped and then
• all executors were stopped (by stopping the containers or killing the processes)

Now you are starting the Executor again, but it keeps on logging

10:34:09.388 [Thread-61] DEBUG o.a.m.l.scheduler.LogstashScheduler - Declined offer......  reason: Host already running task
10:34:09.390 [Thread-61] DEBUG o.a.m.l.scheduler.LogstashScheduler - Declined offer......  reason: Host already running task

No parameters allow recovering from this situation

[jhftrifork]

@sadovnikov can you still reproduce this with 0.8+? I think this is the same issue as #83.

(You will have to clear out ZK after upgrading to 0.8, else the problem will still occur.)