X-Git-Url: https://gerrit.opnfv.org/gerrit/gitweb?a=blobdiff_plain;f=yardstick%2Fbenchmark%2Frunners%2Fiteration.py;h=58ab06a3244f890a1432b01bcfa458aae7513cbe;hb=5f975bd39806cc593686c82c8d5d5975b5342b5f;hp=973bb9ac4030106630d1dccaf7f4688f9ca16784;hpb=e80c35164e7dfee4fe4a3652b71b8775c1c0857a;p=yardstick.git diff --git a/yardstick/benchmark/runners/iteration.py b/yardstick/benchmark/runners/iteration.py index 973bb9ac4..58ab06a32 100644 --- a/yardstick/benchmark/runners/iteration.py +++ b/yardstick/benchmark/runners/iteration.py @@ -20,17 +20,23 @@ """ from __future__ import absolute_import -import os -import multiprocessing + import logging -import traceback +import multiprocessing import time +import traceback + +import os 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): @@ -53,10 +59,6 @@ def _worker_process(queue, cls, method_name, scenario_cfg, method = getattr(benchmark, method_name) - queue.put({'runner_id': runner_cfg['runner_id'], - 'scenario_cfg': scenario_cfg, - 'context_cfg': context_cfg}) - sla_action = None if "sla" in scenario_cfg: sla_action = scenario_cfg["sla"].get("action", "assert") @@ -70,15 +72,17 @@ 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": raise elif sla_action == "monitor": - LOG.warning("SLA validation failed: %s", assertion.args) - errors = assertion.args + LOG.warning("SLA validation failed: %s", error.args) + errors = error.args elif sla_action == "rate-control": try: scenario_cfg['options']['rate'] @@ -89,14 +93,17 @@ def _worker_process(queue, cls, method_name, scenario_cfg, scenario_cfg['options']['rate'] -= delta sequence = 1 continue - except Exception as e: + except Exception: # pylint: disable=broad-except errors = traceback.format_exc() - LOG.exception(e) + LOG.exception("") + raise 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(), @@ -105,10 +112,7 @@ def _worker_process(queue, cls, method_name, scenario_cfg, 'errors': errors } - record = {'runner_id': runner_cfg['runner_id'], - 'benchmark': benchmark_output} - - queue.put(record) + queue.put(benchmark_output, True, QUEUE_PUT_TIMEOUT) LOG.debug("runner=%(runner)s seq=%(sequence)s END", {"runner": runner_cfg["runner_id"], @@ -121,7 +125,18 @@ def _worker_process(queue, cls, method_name, scenario_cfg, LOG.info("worker END") break if "teardown" in run_step: - 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 IterationRunner(base.Runner): @@ -142,7 +157,9 @@ If the scenario ends before the time has elapsed, it will be started again. __execution_type__ = 'Iteration' 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))