LNST-developers March 2021

lnst-developers@lists.fedorahosted.org

5 participants
13 discussions

[PATCH] Revert "RecipeCommon.Perf.Evaluators.BaselineFlowAverageEvaluator: override group_results"
by Jan Tluka 05 Mar '21

05 Mar '21

This reverts commit 0271ddd201033346938ab8dc98d6cf1610e7a2d5. It turned out that this patch is not required. If a user needs to evaluate the aggregated flows a new evaluator class should be created and added to the evaluators list. The aggregation of the data then should be called from within this new evaluator class. Signed-off-by: Jan Tluka <jtluka(a)redhat.com> --- .../Perf/Evaluators/BaselineFlowAverageEvaluator.py | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/lnst/RecipeCommon/Perf/Evaluators/BaselineFlowAverageEvaluator.py b/lnst/RecipeCommon/Perf/Evaluators/BaselineFlowAverageEvaluator.py index 26b6f740..5d2c014d 100644 --- a/lnst/RecipeCommon/Perf/Evaluators/BaselineFlowAverageEvaluator.py +++ b/lnst/RecipeCommon/Perf/Evaluators/BaselineFlowAverageEvaluator.py @@ -30,16 +30,6 @@ class BaselineFlowAverageEvaluator(BaselineEvaluator): "receiver_cpu_stats", ] - def group_results( - self, - recipe: BaseRecipe, - recipe_conf: PerfRecipeConf, - results: List[PerfMeasurementResults], - ) -> List[List[PerfMeasurementResults]]: - new_results = results[0].measurement.aggregate_multi_flow_results(results) - - return [new_results] - def describe_group_results( self, recipe: BaseRecipe, -- 2.26.2

2 1

[PATCH 1/4] Tests.PacketAssert: fix stderr handling
by olichtne＠redhat.com 05 Mar '21

05 Mar '21

From: Ondrej Lichtner <olichtne(a)redhat.com> Tcpdump always prints information to stderr, the "ignore_exprs" regex detection of "real" errors vs debug information is insufficient due to this potentially changing at any point from tcpdump. An example of that is that on RHEL8 we now also see the following line: dropped privs to tcpdump Adjusted the PacketAssert module to only report PASS/FAIL based on the return code of tcpdump (if not 0 then there's some issue). Text in stderr should also always be debug logged, just in case it could be useful later. I also adjusted how the process is "finished" after an interrupt is received so that the returncode is filled in properly (after the communicate call). Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com> --- lnst/Tests/PacketAssert.py | 39 ++++++++++---------------------------- 1 file changed, 10 insertions(+), 29 deletions(-) diff --git a/lnst/Tests/PacketAssert.py b/lnst/Tests/PacketAssert.py index 25a92ba0..75396101 100644 --- a/lnst/Tests/PacketAssert.py +++ b/lnst/Tests/PacketAssert.py @@ -37,25 +37,6 @@ def _check_line(self, line): return self._p_recv += 1 - def _is_real_err(self, err): - - ignore_exprs = [r"tcpdump: verbose output suppressed, use -v or -vv for full protocol decode", - r"listening on %s, link-type .* \(.*\), capture size [0-9]* bytes" % - self.params.interface.name, r"\d+ packets captured", - r"\d+ packets received by filter", r"\d+ packets dropped by kernel"] - - for line in err.split('\n'): - if not line: - continue - match = False - for expr in ignore_exprs: - if re.search(expr, line): - match = True - break - if not match: - return True - return False - def run(self): self._res_data = {} if not is_installed("tcpdump"): @@ -75,17 +56,14 @@ def run(self): except: raise LnstError("Could not handle interrupt properly.") - with packet_assert_process.stdout, packet_assert_process.stderr: - stderr=packet_assert_process.stderr.read().decode() - stdout=packet_assert_process.stdout.read().decode() + stdout, stderr = packet_assert_process.communicate() + stdout = stdout.decode() + stderr = stderr.decode() self._res_data["stderr"] = stderr - - if self._is_real_err(stderr): - self._res_data["msg"] = "errors reported by tcpdump" - logging.error(self._res_data["msg"]) - logging.error(self._res_data["stderr"]) - return False + # tcpdump always reports information to stderr, there may be actual + # errors but also just generic debug information + logging.debug(self._res_data["stderr"]) for line in stdout.split("\n"): self._check_line(line) @@ -93,4 +71,7 @@ def run(self): logging.debug("Capturing finised. Received %d packets." % self._p_recv) self._res_data["p_recv"] = self._p_recv - return True + if packet_assert_process.returncode != 0: + return False + else: + return True -- 2.30.1

4 9

[PATCH v3 00/14] Enable parallel iperf testing
by Jan Tluka 02 Mar '21

02 Mar '21

This patchset is another attempt to implement parallel iperf testing. In comparison to the first RFC the changes are less intrusive and reuses the original Flow concept. Motivation: The current parallel implementation that can be achieved by specifying the perf_parallel_streams recipe parameter works correctly however the limitation is that there's only one iperf process that creates multiple connections and that process (and all the connections) can be handled by a single CPU at the same time. In our internal testing this proved to report very variable CPU utilization numbers. This patchset extends the IperfFlowMeasurementGenerator with additional recipe parameters: * perf_parallel_processes * perf_tool_cpu_policy Additionaly some of the parameters were modified to support parallelism: * perf_tool_cpu is now a ListParam * dev_intr_cpu is now a ListParam The patch set includes also update of DevInterruptHWConfigMixin that is required for this test scenario to provide reproducible results. v2: - made aggregate_multi_flow_results a @staticmethod of BaseFlowMeasurement class - fixed minor issue with perf_tool_cpu_policy parameter handling - removed unused import of PerfFlow in SimpleMacsecRecipe Jan Tluka (14): RecipeCommon.Perf.BaseFlowMeasurement.Flow: make some arguments optional RecipeCommon.Perf.Measurements.BaseFlowMeasurement.Flow: add receiver_port RecipeCommon.Perf.Measurements.IperfFlowMeasurement: configure receiver_port Recipes.ENRT.MeasurementGenerators.IperfMeasurementGenerator: adapt to Flow port changes Recipes.ENRT.ConfigMixins.Reversible: adapt to Flow port changes Recipes.ENRT.MeasurementGenerators.IperfMeasurementGenerator: add _create_perf_flows Recipes.ENRT.MeasurementGenerators.IperfMeasurementGenerator: add perf_parallel_processes parameter IperfMeasurementGenerator: adjust cpu parameters for parallel iperf support RecipeCommon.Perf.Measurements.IperfFlowMeasurement: adjust cpupin for both server and client RecipeCommon.Perf.Measurements.BaseFlowMeasurement: add aggregate_multi_flow_results() RecipeCommon.Perf.Evaluators.BaselineFlowAverageEvaluator: override group_results RecipeCommon.Perf.Measurements.BaseFlowMeasurement: report also aggregated results Recipes.ENRT.ConfigMixins.DevInterruptHWConfigMixin: change dev_intr_cpu to ListParam Recipes.ENRT.SimpleMacsecRecipe: remove unused import .../BaselineFlowAverageEvaluator.py | 10 +++ .../Perf/Measurements/BaseFlowMeasurement.py | 74 +++++++++++++++- .../Perf/Measurements/IperfFlowMeasurement.py | 38 ++++---- .../ConfigMixins/DevInterruptHWConfigMixin.py | 26 +++--- .../ConfigMixins/PerfReversibleFlowMixin.py | 34 +++++-- .../IperfMeasurementGenerator.py | 88 +++++++++++++++---- lnst/Recipes/ENRT/SimpleMacsecRecipe.py | 1 - 7 files changed, 217 insertions(+), 54 deletions(-) -- 2.26.2

2 15

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

LNST-developers March 2021