Fix indices in initial task-to-task registration

[EnricoMi]

In Elastic Spark, the initial task-to-task registration has to use the actual next index, not the index + 1, as in elastic mode not all indices might exist on initialisation.

Fixes these errors:

Exception in thread Thread-18:
Traceback (most recent call last):
  File "/usr/lib/python3.7/threading.py", line 926, in _bootstrap_inner
    self.run()
  File "/usr/lib/python3.7/threading.py", line 870, in run
    self._target(*self._args, **self._kwargs)
  File "/usr/local/lib/python3.7/dist-packages/horovod/spark/runner.py", line 179, in notify_and_register
    next_task_addresses = driver.all_task_addresses(next_task_index)
  File "/usr/local/lib/python3.7/dist-packages/horovod/runner/common/service/driver_service.py", line 112, in all_task_addresses
    return self._all_task_addresses[index].copy()
KeyError: 2

[github-actions]

Unit Test Results

     773 files  ±0       773 suites  ±0   8h 49m 38s ⏱️ - 17m 21s
     722 tests ±0       661 ✔️  - 14       47 💤 ±0  14 ❌ +14 
16 693 runs  ±0  11 761 ✔️  - 15  4 917 💤 ±0  15 ❌ +15 

For more details on these failures, see this check.

Results for commit 5f72273. ± Comparison against base commit 046c071.

[github-actions]

Unit Test Results (with flaky tests)

     963 files  +     88       963 suites  +88   10h 33m 37s ⏱️ + 29m 52s
     722 tests ±       0       657 ✔️  -      16       47 💤 ±    0  18 ❌ +16 
21 036 runs  +1 972  14 660 ✔️ +1 336  6 325 💤 +588  51 ❌ +48 

For more details on these failures, see this check.

Results for commit 5f72273. ± Comparison against base commit 046c071.