Specifications

6
Troubleshooting
111
Responding to specific alarms
Responding to a communication lost alarm
TheShelfManagerisnotgettingkeepaliveresponsesfromthetargetFRU ,usuallybecause
theFRUwasextractedwithoutgoingthroughthenormalhotswapextractionprocedure.Try
thesesteps:
•IftheFRUhasbeenremoved,theFRUcanberemovedfromtheRPTusingHPIcontrol
0x101eonthe
ShelfManagerresource,theFailedResourceExtractcontrol.Thisalso
removesthealarm.SeetheSAFMappingSpecificationfordetails.
InitiateadiscoveryoperationfromtheplatformmanagementCLIusingthe
rediscoverShelfcommand.Wait60secondsafterthediscoverycompletestoseeifthe
alarmisremoved.
•IftheFRUisstillintheshelf,physicallyextractandreinserttheFRUaccordingtothe
procedureforremovingandreplacingamoduleinthePlatformReferencemanual.
•Iftheconditionstillpersists,afirmwarelockuporotherhardwareproblemmayhave
occurredintheIPMCorMMC.However,thefullsetofFRUstowhichcommunicationhas
beenlostshouldbeanalyzedforpatterns.Theanalysismayindicateaproblemaffecting
multipleFRUsoraffectingahubmodule’sabilitytocommunicatewithallotherFRUs.
Iftheconditionpersists,contactRadisysTechnicalSupportforassistance.
Responding to an IPMB disabled alarm
TheinterfacebetweentheIPMCandtheexternalbushasbeendisabledbytheIPMC.
1. IfaFRUlosescontactwiththeIPMB,querytheFRU’sIPMBsensortodeterminethe
reason.
2. IftheentireIPMBisdown,theShelfManagermodulegeneratesanalarm.Querythe
module’ssensortodeterminethereason.
TointerpretthevaluesofthephysicalIPMB0sensors,seetheAdvancedTCABase
Specification.Iftheconditionpersists,contactRadisysTechnicalSupportforassistance.
Responding to a Resource Failed alarm
Thefollowingsymptomsapplytothisalarm:
•AFRUenterstheResourceFailedstateandaResourceFailedalarmisaddedtotheDomain
AlarmTable.
•TheFRUmaybeperiodicallyenteringandleavingtheResourceFailedstateaswell.