,Benefitingfromitsexcellentperformanceandscalability,Cephdistributedstorageisalsoconfrontingtheproblems,i.e.theunnecessarytransferofthedataleadingtotheincreasingconsumptionofthesystemwhichistriggeredbytheadditionsanddeletionsontheequipment.AimingattheimplementmethodofCephinsystemalgorithmandlogicallayout,thispaperhascarriedonresearch,analyzingitsdefectsintheaspectofdatatransferandresourceconsumptionandputsforwardthehandlingmethodofthefailurenodewhentheclusterstoragedevicefails.Throughtheestablishmentoftheclusterdeviceflagintheprocessofdatatransfer,theoptimizedschemeofthedatamigrationintheenvironmentofproductionisrealized,andtheutilizationofthesystemresourcesisimproved.Theexperimentaltestresultshaveshownthattheschemecanreduceaboutthe30%-40%oftransfervolume,effectivelylowertheresourcesconsumptionofCephdistributedstorage,aswellaspreventinvalidandexcessivedatatransfer.
2017/6/2 11:03:30
432KB
ceph
1