[
https://issues.apache.org/jira/browse/BEAM-5985?focusedWorklogId=208796&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-208796
]
ASF GitHub Bot logged work on BEAM-5985:
----------------------------------------
Author: ASF GitHub Bot
Created on: 06/Mar/19 13:21
Start Date: 06/Mar/19 13:21
Worklog Time Spent: 10m
Work Description: kkucharc commented on pull request #7903: [BEAM-5985]
Dataflow batch load test jobs
URL: https://github.com/apache/beam/pull/7903#discussion_r262937050
##########
File path: .test-infra/jenkins/job_LoadTests_Java.groovy
##########
@@ -17,118 +17,213 @@
*/
import CommonJobProperties as commonJobProperties
+import CommonTestProperties
import LoadTestsBuilder as loadTestsBuilder
import PhraseTriggeringPostCommitBuilder
+import CronJobBuilder
def loadTestConfigurations = [
[
- jobName :
'beam_Java_LoadTests_GroupByKey_Dataflow_Small',
- jobDescription : 'Runs GroupByKey load tests on Dataflow
runner small records 10b',
- itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
- prCommitStatusName: 'Java GroupByKey Small Load Test Dataflow',
- prTriggerPhrase : 'Run GroupByKey Small Java Load Test
Dataflow',
- runner : CommonTestProperties.Runner.DATAFLOW,
+ title : 'Load test: 2GB of 10B records',
+ itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
+ runner : CommonTestProperties.Runner.DATAFLOW,
jobProperties : [
project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_1',
tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
publishToBigQuery : true,
- bigQueryDataset : 'load_test_PRs',
- bigQueryTable : 'dataflow_gbk_small',
- sourceOptions :
'{"numRecords":1000000000,"splitPointFrequencyRecords":1,"keySizeBytes":1,"valueSizeBytes":9,"numHotKeys":0,"hotKeyFraction":0,"seed":123456,"bundleSizeDistribution":{"type":"const","const":42},"forceNumInitialBundles":100,"progressShape":"LINEAR","initializeDelayDistribution":{"type":"const","const":42}}',
- stepOptions :
'{"outputRecordsPerInputRecord":1,"preservesInputKeyDistribution":true,"perBundleDelay":10000,"perBundleDelayType":"MIXED","cpuUtilizationInMixedDelay":0.5}',
- fanout : 10,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_1',
+ sourceOptions : """
+ {
+ "numRecords": 200000000,
+ "keySizeBytes": 1,
+ "valueSizeBytes": 9
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 1,
iterations : 1,
- maxNumWorkers : 32,
+ maxNumWorkers : 5,
+ numWorkers : 5,
+ autoscalingAlgorithm: "NONE"
]
-
],
-]
-
-for (testConfiguration in loadTestConfigurations) {
- PhraseTriggeringPostCommitBuilder.postCommitJob(
- testConfiguration.jobName,
- testConfiguration.prTriggerPhrase,
- testConfiguration.prCommitStatusName,
- this
- ) {
- description(testConfiguration.jobDescription)
- commonJobProperties.setTopLevelMainJobProperties(delegate, 'master',
240)
- loadTestsBuilder.loadTest(delegate, testConfiguration.jobDescription,
testConfiguration.runner, testConfiguration.jobProperties,
testConfiguration.itClass)
- }
-}
-
-def smokeTestConfigurations = [
[
- title : 'GroupByKey load test Direct',
+ title : 'Load test: 2GB of 100B records',
itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
- runner : CommonTestProperties.Runner.DIRECT,
- jobProperties: [
- publishToBigQuery: true,
- bigQueryDataset : 'load_test_SMOKE',
- bigQueryTable : 'direct_gbk',
- sourceOptions :
'{"numRecords":100000,"splitPointFrequencyRecords":1}',
- stepOptions :
'{"outputRecordsPerInputRecord":1,"preservesInputKeyDistribution":true}',
- fanout : 10,
- iterations : 1,
+ runner : CommonTestProperties.Runner.DATAFLOW,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_2',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_2',
+ sourceOptions : """
+ {
+ "numRecords": 20000000,
+ "keySizeBytes": 10,
+ "valueSizeBytes": 90
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 1,
+ iterations : 1,
+ maxNumWorkers : 5,
+ numWorkers : 5,
+ autoscalingAlgorithm: "NONE"
]
],
[
- title : 'GroupByKey load test Dataflow',
+ title : 'Load test: 2GB of 100kB records',
itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
runner : CommonTestProperties.Runner.DATAFLOW,
- jobProperties: [
- project : 'apache-beam-testing',
- tempLocation :
'gs://temp-storage-for-perf-tests/smoketests',
- publishToBigQuery: true,
- bigQueryDataset : 'load_test_SMOKE',
- bigQueryTable : 'dataflow_gbk',
- sourceOptions :
'{"numRecords":100000,"splitPointFrequencyRecords":1}',
- stepOptions :
'{"outputRecordsPerInputRecord":1,"preservesInputKeyDistribution":true}',
- fanout : 10,
- iterations : 1,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_3',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_3',
+ sourceOptions : """
+ {
+ "numRecords": 2000,
+ "keySizeBytes": 100000,
+ "valueSizeBytes": 900000
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 1,
+ iterations : 1,
+ maxNumWorkers : 5,
+ numWorkers : 5,
+ autoscalingAlgorithm: "NONE"
]
],
[
- title : 'GroupByKey load test Flink',
+ title : 'Load test: fanout 4 times with 2GB 10-byte
records total',
itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
- runner : CommonTestProperties.Runner.FLINK,
- jobProperties: [
- publishToBigQuery: true,
- bigQueryDataset : 'load_test_SMOKE',
- bigQueryTable : 'flink_gbk',
- sourceOptions :
'{"numRecords":100000,"splitPointFrequencyRecords":1}',
- stepOptions :
'{"outputRecordsPerInputRecord":1,"preservesInputKeyDistribution":true}',
- fanout : 10,
- iterations : 1,
+ runner : CommonTestProperties.Runner.DATAFLOW,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_4',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_4',
+ sourceOptions : """
+ {
+ "numRecords": 5000000,
+ "keySizeBytes": 10,
+ "valueSizeBytes": 90
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 4,
+ iterations : 1,
+ maxNumWorkers : 16,
+ numWorkers : 16,
+ autoscalingAlgorithm: "NONE"
]
],
[
- title : 'GroupByKey load test Spark',
+ title : 'Load test: fanout 8 times with 2GB 10-byte
records total',
itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
- runner : CommonTestProperties.Runner.SPARK,
- jobProperties: [
- sparkMaster : 'local[4]',
- publishToBigQuery: true,
- bigQueryDataset : 'load_test_SMOKE',
- bigQueryTable : 'spark_gbk',
- sourceOptions :
'{"numRecords":100000,"splitPointFrequencyRecords":1}',
- stepOptions :
'{"outputRecordsPerInputRecord":1,"preservesInputKeyDistribution":true}',
- fanout : 10,
- iterations : 1,
+ runner : CommonTestProperties.Runner.DATAFLOW,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_5',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_5',
+ sourceOptions : """
+ {
+ "numRecords": 2500000,
+ "keySizeBytes": 10,
+ "valueSizeBytes": 90
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 8,
+ iterations : 1,
+ maxNumWorkers : 16,
+ numWorkers : 16,
+ autoscalingAlgorithm: "NONE"
+ ]
+ ],
+ [
+ title : 'Load test: reiterate 4 times 10kB values',
+ itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
+ runner : CommonTestProperties.Runner.DATAFLOW,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_6',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_6',
+ sourceOptions : """
+ {
+ "numRecords": 20000000,
+ "keySizeBytes": 10,
+ "valueSizeBytes": 90,
+ "numHotKeys": 200,
+ "hotKeyFraction": 1
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 1,
+ iterations : 4,
+ maxNumWorkers : 5,
+ numWorkers : 5,
+ autoscalingAlgorithm: "NONE"
+ ]
+ ],
+ [
+ title : 'Load test: reiterate 4 times 2MB values',
+ itClass :
'org.apache.beam.sdk.loadtests.GroupByKeyLoadTest',
+ runner : CommonTestProperties.Runner.DATAFLOW,
+ jobProperties : [
+ project : 'apache-beam-testing',
+ appName :
'load_tests_Java_Dataflow_Batch_GBK_7',
+ tempLocation :
'gs://temp-storage-for-perf-tests/loadtests',
+ publishToBigQuery : true,
+ bigQueryDataset : 'load_test',
+ bigQueryTable : 'java_dataflow_batch_GBK_7',
+ sourceOptions : """
+ {
+ "numRecords": 20000000,
+ "keySizeBytes": 10,
+ "valueSizeBytes": 90,
+ "numHotKeys": 10,
+ "hotKeyFraction": 1
+ }
+ """.trim().replaceAll("\\s", ""),
+ fanout : 1,
+ iterations : 4,
+ maxNumWorkers : 5,
+ numWorkers : 5,
+ autoscalingAlgorithm: "NONE"
]
]
]
+def loadTestJob = { scope ->
+ scope.description('Runs Java GBK load tests on Dataflow runner in batch
mode')
+ commonJobProperties.setTopLevelMainJobProperties(scope, 'master', 240)
+
+ for (testConfiguration in loadTestConfigurations) {
+ loadTestsBuilder.loadTest(scope, testConfiguration.title,
testConfiguration.runner, testConfiguration.jobProperties,
testConfiguration.itClass, CommonTestProperties.TriggeringContext.PR)
Review comment:
Just to remember: merge to master and add `SDK` enum parameter in tests.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 208796)
Time Spent: 15.5h (was: 15h 20m)
> Create jenkins jobs to run the load tests for Java SDK
> ------------------------------------------------------
>
> Key: BEAM-5985
> URL: https://issues.apache.org/jira/browse/BEAM-5985
> Project: Beam
> Issue Type: Sub-task
> Components: testing
> Reporter: Lukasz Gajowy
> Assignee: Kasia Kucharczyk
> Priority: Major
> Time Spent: 15.5h
> Remaining Estimate: 0h
>
> How/how often/in what cases we run those tests is yet to be decided (this is
> part of the task)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
