You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "Łukasz Gajowy (JIRA)" <ji...@apache.org> on 2018/05/16 12:59:00 UTC
[jira] [Updated] (BEAM-4140) Beam repo is cloned 2 times instead of
one during Performance test jenkins Jobs
[ https://issues.apache.org/jira/browse/BEAM-4140?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Łukasz Gajowy updated BEAM-4140:
--------------------------------
Description:
Example jenkins job logs can be found here (any job): [https://builds.apache.org/view/A-D/view/Beam/job/beam_PerformanceTests_JDBC/] . This is best visible if one searches logs by "git clone".
First clone is done by a jenkins job to a $WORKSPACE/src directory:
[https://github.com/apache/beam/blob/9c9f4ceceb87933da2b03304efd21edf55216937/.test-infra/jenkins/common_job_properties.groovy#L96]
Second one is performed by Perfkit:
[https://github.com/GoogleCloudPlatform/PerfKitBenchmarker/blob/f852afab0f71489958240299706878c1e51f1e89/perfkitbenchmarker/beam_benchmark_helper.py#L141]
There are two reasons I think this should be changed:
1. This seems to be redundant. We either can utilize "beam_location" perfkit's flag or not clone the repo to '$WORKSPACE/src' directory.
2. By cloning inside Perfkit we use the master branch. This is misguiding when users want to test their Pull Requests in terms of performance (this is possible thanks to current Jenkins job configuration).
was:
Example jenkins job logs can be found here (any job): [https://builds.apache.org/view/A-D/view/Beam/job/beam_PerformanceTests_JDBC/] . This is best visible if one searches logs by "git clone".
First clone is done by a jenkins job to a $WORKSPACE/src directory:
[https://github.com/apache/beam/blob/9c9f4ceceb87933da2b03304efd21edf55216937/.test-infra/jenkins/common_job_properties.groovy#L96]
Second one is performed by Perfkit:
[https://github.com/GoogleCloudPlatform/PerfKitBenchmarker/blob/f852afab0f71489958240299706878c1e51f1e89/perfkitbenchmarker/beam_benchmark_helper.py#L141]
There are two reasons I think this should be changed:
1. This seems to be redundant. We either can utilize "beam_location" perfkit's flag or not clone the repo to '$WORKSPACE/src' directory.
2. By cloning inside Perfkit we use the master branch. This is misguiding when users want to test their Pull Requests in terms of performance (this is possible thanks to current Jenkins job configuration
> Beam repo is cloned 2 times instead of one during Performance test jenkins Jobs
> -------------------------------------------------------------------------------
>
> Key: BEAM-4140
> URL: https://issues.apache.org/jira/browse/BEAM-4140
> Project: Beam
> Issue Type: Bug
> Components: testing
> Reporter: Łukasz Gajowy
> Assignee: Łukasz Gajowy
> Priority: Minor
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Example jenkins job logs can be found here (any job): [https://builds.apache.org/view/A-D/view/Beam/job/beam_PerformanceTests_JDBC/] . This is best visible if one searches logs by "git clone".
> First clone is done by a jenkins job to a $WORKSPACE/src directory:
> [https://github.com/apache/beam/blob/9c9f4ceceb87933da2b03304efd21edf55216937/.test-infra/jenkins/common_job_properties.groovy#L96]
> Second one is performed by Perfkit:
> [https://github.com/GoogleCloudPlatform/PerfKitBenchmarker/blob/f852afab0f71489958240299706878c1e51f1e89/perfkitbenchmarker/beam_benchmark_helper.py#L141]
> There are two reasons I think this should be changed:
> 1. This seems to be redundant. We either can utilize "beam_location" perfkit's flag or not clone the repo to '$WORKSPACE/src' directory.
> 2. By cloning inside Perfkit we use the master branch. This is misguiding when users want to test their Pull Requests in terms of performance (this is possible thanks to current Jenkins job configuration).
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)