X-Git-Url: https://gerrit.opnfv.org/gerrit/gitweb?a=blobdiff_plain;f=yardstick%2Fbenchmark%2Frunners%2Fduration.py;h=14fd8bb4717a6efa9d18bfa7d6111d76ec907558;hb=60e326463ae06038facf79f14e29901e9596e0dc;hp=c2c6a8f19932fc4a2bb076d373666254aaba5cd9;hpb=75ff03570c08093fbe5a2bb85856ec39752f900d;p=yardstick.git

diff --git a/yardstick/benchmark/runners/duration.py b/yardstick/benchmark/runners/duration.py
index c2c6a8f19..14fd8bb47 100644
--- a/yardstick/benchmark/runners/duration.py
+++ b/yardstick/benchmark/runners/duration.py
@@ -27,10 +27,14 @@ import traceback
 import time
 
 from yardstick.benchmark.runners import base
+from yardstick.common import exceptions as y_exc
 
 LOG = logging.getLogger(__name__)
 
 
+QUEUE_PUT_TIMEOUT = 10
+
+
 def _worker_process(queue, cls, method_name, scenario_cfg,
                     context_cfg, aborted, output_queue):
 
@@ -54,6 +58,7 @@ def _worker_process(queue, cls, method_name, scenario_cfg,
         sla_action = scenario_cfg["sla"].get("action", "assert")
 
     start = time.time()
+    timeout = start + duration
     while True:
 
         LOG.debug("runner=%(runner)s seq=%(sequence)s START",
@@ -62,23 +67,30 @@ def _worker_process(queue, cls, method_name, scenario_cfg,
         data = {}
         errors = ""
 
+        benchmark.pre_run_wait_time(interval)
+
         try:
             result = method(data)
-        except AssertionError as assertion:
+        except y_exc.SLAValidationError as error:
             # SLA validation failed in scenario, determine what to do now
             if sla_action == "assert":
+                benchmark.teardown()
                 raise
             elif sla_action == "monitor":
-                LOG.warning("SLA validation failed: %s", assertion.args)
-                errors = assertion.args
-        except Exception as e:
+                LOG.warning("SLA validation failed: %s", error.args)
+                errors = error.args
+        # catch all exceptions because with multiprocessing we can have un-picklable exception
+        # problems  https://bugs.python.org/issue9400
+        except Exception:  # pylint: disable=broad-except
             errors = traceback.format_exc()
-            LOG.exception(e)
+            LOG.exception("")
         else:
             if result:
-                output_queue.put(result)
+                # add timeout for put so we don't block test
+                # if we do timeout we don't care about dropping individual KPIs
+                output_queue.put(result, True, QUEUE_PUT_TIMEOUT)
 
-        time.sleep(interval)
+        benchmark.post_run_wait_time(interval)
 
         benchmark_output = {
             'timestamp': time.time(),
@@ -87,19 +99,29 @@ def _worker_process(queue, cls, method_name, scenario_cfg,
             'errors': errors
         }
 
-        queue.put(benchmark_output)
+        queue.put(benchmark_output, True, QUEUE_PUT_TIMEOUT)
 
         LOG.debug("runner=%(runner)s seq=%(sequence)s END",
                   {"runner": runner_cfg["runner_id"], "sequence": sequence})
 
         sequence += 1
 
-        if (errors and sla_action is None) or \
-                (time.time() - start > duration or aborted.is_set()):
+        if (errors and sla_action is None) or time.time() > timeout or aborted.is_set():
             LOG.info("Worker END")
             break
 
-    benchmark.teardown()
+    try:
+        benchmark.teardown()
+    except Exception:
+        # catch any exception in teardown and convert to simple exception
+        # never pass exceptions back to multiprocessing, because some exceptions can
+        # be unpicklable
+        # https://bugs.python.org/issue9400
+        LOG.exception("")
+        raise SystemExit(1)
+
+    LOG.debug("queue.qsize() = %s", queue.qsize())
+    LOG.debug("output_queue.qsize() = %s", output_queue.qsize())
 
 
 class DurationRunner(base.Runner):
@@ -120,7 +142,9 @@ If the scenario ends before the time has elapsed, it will be started again.
     __execution_type__ = 'Duration'
 
     def _run_benchmark(self, cls, method, scenario_cfg, context_cfg):
+        name = "{}-{}-{}".format(self.__execution_type__, scenario_cfg.get("type"), os.getpid())
         self.process = multiprocessing.Process(
+            name=name,
             target=_worker_process,
             args=(self.result_queue, cls, method, scenario_cfg,
                   context_cfg, self.aborted, self.output_queue))