Chukwa A large-scale monitoring system

Chukwa A large-scale monitoring system

ID:39776144

大小:155.58 KB

页数:5页

时间:2019-07-11

Chukwa A large-scale monitoring system_第1页
Chukwa A large-scale monitoring system_第2页
Chukwa A large-scale monitoring system_第3页
Chukwa A large-scale monitoring system_第4页
Chukwa A large-scale monitoring system_第5页
资源描述:

《Chukwa A large-scale monitoring system》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、Chukwa:Alarge-scalemonitoringsystemJeromeBoulonAndyKonwinskiRunpingQijboulon@yahoo-inc.comandyk@cs.berkeley.edurunping@yahoo-inc.comYahoo!,incUCBerkeleyYahoo!,incArielRabkinEricYangMacYangasrabkin@cs.berkeley.edueyang@yahoo-inc.commacyang@yahoo-inc.comUCBerkeleyYahoo!,incYahoo!,incAbs

2、tractdata.Logsareperiodicallydeleted,unlessuserstaketheinitiativeinstoringthem.WedescribethedesignandinitialimplementationofWebelievethatoursituationistypical,andthatlo-Chukwa,adatacollectionsystemformonitoringandan-calstorageofloggingdataisacommonmodelforveryalyzinglargedistributedsy

3、stems.Chukwaisbuiltonlargedeployments.TotheextentthatmoresophisticatedtopofHadoop,anopensourcedistributedfilesystemanddatamanagementtechniquesareutilized,theyarelargelyMapReduceimplementation,andinheritsHadoop’sscal-supportedbyad-hocproprietarysolutions.Awelldocu-abilityandrobustness.C

4、hukwaalsoincludesaflexiblementedopensourcetoolsetforhandlingmonitoringdataandpowerfultoolkitfordisplayingmonitoringandanal-thussolvesasignificantpracticalproblemandprovidesysisresults,inordertomakethebestuseofthiscollectedavaluablereferencepointforfuturedevelopmentinthisdata.area.Wedidn

5、otaimtosolvetheproblemofreal-timemon-1Introductionitoringforfailuredetection,whichsystemssuchasGan-gliaalreadydowell.Rather,wewantedasystemthatHadoopisadistributedfilesystemandMapReduce[1]wouldprocesslargevolumesofdatainatimescaleofimplementationthatisusedpervasivelyatYahoo!foraminutes

6、,notseconds,todetectmoresubtleconditions,varietyofcriticalbusinesspurposes.Productionclustersandtoaidinfailurediagnosis.Humanengineersandop-oftenincludethousandsofnodes.Largedistributedsys-eratorsdonotgenerallyreactonatimescaleofseconds,temssuchasHadooparefearsomelycomplex,andcanandso

7、aprocessingdelayofafewminutesisnotacon-failincomplicatedandsubtleways.Asaresult,Hadoopcernforus.isextensivelyinstrumented.Atwo-thousandnodeclus-Weareintheprocessofbuildingasystem,whichweterconfiguredfornormaloperationgeneratesnearlyhalfcallChukwa,todemonstratethatpracticallarge-scaleca

8、nbere

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。