vaidya简介
- hadoop 0.20.2的vaidya文档;http://hadoop.apache.org/docs/r0.20.2/vaidya.html
- 一个vaidya使用的简介;http://www.cnblogs.com/reckzhou/articles/2409577.html
- 其他hadoop监控工具http://www.zhihu.com/question/19661847 http://www.zhihu.com/question/19661847
- 使用示例
vaidya实现细节与内置的测试类
测试用例的配置文件
<DiagnosticTest>
<Title><![CDATA[Balanaced Reduce Partitioning]]></Title> 直接复制到结果Title中
<ClassName><![CDATA[org.apache.hadoop.vaidya.postexdiagnosis.tests.BalancedReducePartitioning]]></ClassName> 调用的Test类,调用这个类的run函数
<Description><![CDATA[This rule tests as to how well the input to reduce tasks is balanced]]></Description> 直接复制到结果的Description中
<Importance><![CDATA[High]]></Importance> 直接复制到结果的Importance中
<SuccessThreshold><![CDATA[
0.20
]]></SuccessThreshold>,
<Prescription><![CDATA[advice]]></Prescription>
<InputElement>
<PercentReduceRecords><![CDATA[
0.85
]]></PercentReduceRecords>
</InputElement>
</DiagnosticTest
|
生成结果
- <TestReportElement>
<TestTitle>Balanaced Reduce Partitioning</TestTitle>
<TestDescription>This rule tests as to how well the input to reduce tasks is balanced</TestDescription>
<TestImportance>HIGH</TestImportance>
<TestResult>NEGATIVE(PASSED)</TestResult>
if
impact level > evaluate()接口产生的impact level
<TestSeverity>
0.0
</TestSeverity> evaluate()接口产生的impact level
<ReferenceDetails>* TotalReduceTasks:
1
* BusyReduceTasks processing
0.85
% of total records:
1
* Impact:
0.0
</ReferenceDetails>
利用数据生成的格式化细节
<TestPrescription>* Use the appropriate partitioning function * For streaming job consider following partitioner and hadoop config parameters * org.apache.hadoop.mapred.lib.KeyFieldBasedPartitioner * -jobconf stream.map.output.field.separator, -jobconf stream.num.map.output.key.fields</TestPrescription>
利用数据生成的格式化建议
</TestReportElement>
|
类图
流程图
内置测试类简介
- BalancedReducePartitioning
- MapsReExecutionImpact
- ReadingHDFSFilesAsSideEffect
- MapSideDiskSpill