diff --git a/tests/test_concurrent_executor.py b/tests/test_concurrent_executor.py
index 667e6c3eaf..138bd84925 100644
--- a/tests/test_concurrent_executor.py
+++ b/tests/test_concurrent_executor.py
@@ -7,29 +7,35 @@
 import pytest
 
 DATA = [1, 2, 3, 4]
+RETRIES = 5
 
 
 def _echo(x, please_fail=False):
     if please_fail or x == 'please fail':
-        raise IOError('Fake I/O error, cause you asked')
+        raise IOError('Fake I/O error, because you asked')
     return x
 
 
-def run_tests_for_runner(runner, sleep_amount=0.5):
+def run_executor_tests(executor, sleep_time=1):
     # get_ready: mostly pending
-    futures = runner.map(_echo, DATA)
+    futures = executor.map(_echo, DATA)
     assert len(futures) == len(DATA)
-    completed, failed, pending = runner.get_ready(futures)
+    completed, failed, pending = executor.get_ready(futures)
     assert len(failed) == 0
     assert len(completed) + len(pending) == len(DATA)
 
     # get_ready: processed + failure
     data = ['please fail'] + DATA
-    futures = runner.map(_echo, data)
+    futures = executor.map(_echo, data)
     assert len(futures) == len(data)
-    sleep(sleep_amount)  # give it "enough" time to finish
-    completed, failed, pending = runner.get_ready(futures)
-    if sleep_amount:
+
+    for _ in range(RETRIES):
+        sleep(sleep_time)  # give it "enough" time to finish
+        completed, failed, pending = executor.get_ready(futures)
+        if len(pending) == 0:
+            break
+
+    if sleep_time:
         assert len(failed) == 1
     else:
         assert len(failed) in [0, 1]
@@ -37,46 +43,46 @@ def run_tests_for_runner(runner, sleep_amount=0.5):
     assert len(completed) + len(pending) + len(failed) == len(data)
 
     # test results
-    futures = runner.map(_echo, DATA)
-    results = runner.results(futures)
+    futures = executor.map(_echo, DATA)
+    results = executor.results(futures)
 
     assert len(results) == len(DATA)
     assert set(results) == set(DATA)
 
     # Test failure pass-through
-    future = runner.submit(_echo, "", please_fail=True)
+    future = executor.submit(_echo, "", please_fail=True)
 
-    for ff in runner.as_completed([future]):
+    for ff in executor.as_completed([future]):
         with pytest.raises(IOError):
-            runner.result(ff)
+            executor.result(ff)
 
     # Next completed with data
-    future = runner.submit(_echo, 'tt')
+    future = executor.submit(_echo, 'tt')
     futures = [future]
-    result, futures = runner.next_completed(futures, 'default')
+    result, futures = executor.next_completed(futures, 'default')
     assert len(futures) == 0
     print(type(result), result)
 
     # Next completed with empty list
-    result, futures = runner.next_completed([], 'default')
+    result, futures = executor.next_completed([], 'default')
     assert result == 'default'
     assert len(futures) == 0
 
-    runner.release(future)
+    executor.release(future)
 
 
 def test_concurrent_executor():
-    runner = get_executor(None, 2)
-    assert str(runner).find('Multi') >= 0
-    run_tests_for_runner(runner, 0.3)
+    executor = get_executor(None, 2)
+    assert 'Multiproc' in str(executor)
+    run_executor_tests(executor)
 
-    runner = get_executor(None, 2, use_cloud_pickle=False)
-    assert str(runner).find('Multi') >= 0
-    run_tests_for_runner(runner, 0.3)
+    executor = get_executor(None, 2, use_cloud_pickle=False)
+    assert 'Multiproc' in str(executor)
+    run_executor_tests(executor)
 
 
 def test_fallback_executor():
-    runner = get_executor(None, None)
-    assert str(runner).find('Serial') >= 0
+    executor = get_executor(None, None)
+    assert 'Serial' in str(executor)
 
-    run_tests_for_runner(runner, 0)
+    run_executor_tests(executor, sleep_time=0)