ReferenceSHELF MANAGEMENT SOFTWARESOFTWARE VERSION 4.2.xApril 2012 007-03261-0008
1Chapter10Shelf Management OverviewTheShelfManagerisresponsibleformonitoringconditionsofmodulesandothershelfcomponentsandcontrollingth
6Troubleshooting100General troubleshooting proceduresThissectionexplainshowtoacknowledgeandviewalarms,viewsystemevents,andchecksensoran
6Troubleshooting101Viewing the system event logYoucanviewamodule’slocalSELortheshelf’sSELusingtheopen‐sourceipmitoolutilityorthersys
6Troubleshooting102Investigating sensor valuesToviewthesensorvalueforasensorthathasgeneratedanalarm:1. WritedowntheFRUslotidentified
6Troubleshooting1032. ListthesensorsfortheSPMresource:qsb<resourceId>3. Notetheintaketemperatures.4. Repeatsteps1,2,and3fortheo
6Troubleshooting104On the ATCA-6014 and ATCA-6016 shelves:Eachfantrayhasanexhausttemperaturesensor.Checktheexhausttemperaturesasfollows:1
6Troubleshooting105IPMI frame structure and message flowTherearenoframesinIPMBmessages.Figure 10showsthemessageflowthroughanSCMpairfor
6Troubleshooting106OnlytheswitchingportionoftheSCMsactinActive‐Activemode.TheShMCsareneverindualactivestateunlessthere'sac
6Troubleshooting107EachIPMCthatisnottheactiveShMCmonitorsits“ownside”ofthebus.Inotherwords,eachIPMCismonitoringitsownI2Cinter
6Troubleshooting108TheusercandotoseveralthingsrespondtoapersistentIPMBisolationissuewhenthereisashelfwideIPMBisolationalarmthat
6Troubleshooting109d. ToreturntheboardtoitsdefaultstatewhereitisjoinedonthatIPMByoucanagainusetheCLIasfollows:platform‐mgmt(fron
1Shelf Management Overview11Event and error managementTheShelfManagermanagesnormaleventsanderrorconditionsintheseways:•Powersdevicesupan
6Troubleshooting110Troubleshooting FRU and shelf operation issuesThissectionexplainshowtorespondtoFRUandshelfoperationalarmsandotherissu
6Troubleshooting111Responding to specific alarmsResponding to a communication lost alarmTheShelfManagerisnotgettingkeep‐aliveresponsesfromthe
6Troubleshooting112TherearenohardwarepresencedetectlinesinATCAbackplanes.TheShMSthusreliesonmessagebasedpings toperiodicallyquery
6Troubleshooting113Aftermakinganyshmgr.conffileupdate,remembertosynchronizeittothestandbySCMfromtheplatform‐managementCLIbeforerebo
6Troubleshooting114Responding to an IPMB isolated alarmThemodule’sIPMChasdetectedaproblemononeorbothofitsIPMBinterfaces,andhasisolate
6Troubleshooting1155. Disableeventgenera tiononthatsensorwiththesecommands:8<resourceId><sensorId>Thescreendisplaysthefinalme
6Troubleshooting1165. Lookforthe“inetaddr”value.TheSCMinthelower‐numberedslotinthechassisshouldhaveaneth1IPaddressof10.0.1.1.Th
6Troubleshooting117Beforeaddressingthepossiblecauseandtryingtocorrectit,someoftheinformationfromtheINSERTION_PENDINGstatemaybeusef
6Troubleshooting1183. IftheFRUisfoundtobeinM2andAutoInsertiontimerisnot‐1oralargevalue,checkthe stateofthe'FRUActivation
6Troubleshooting119c. Checkthe'DesiredPower'controlsonallFRUsinstalledonthatslot(parentandmanagedFRUs)andcheckifthetotal
1Shelf Management Overview12Shelf Manager complianceAsofthismanual’spublicationdate,theRadisysShelfManagerisfullycompliantwiththePICMG
6Troubleshooting120Responding to a fan not running1. CheckthattheShelfManagercancommunicatewiththefanandthatitisintheactivestate.Use
6Troubleshooting121Troubleshooting Shelf Manager issuesGeneral issuesWhenyouencounterproblemsrelatedtoshelfmanagement,checkthefollowingitem
6Troubleshooting122Make sure the Shelf Manager software is runningMakesuretheShelfManagersoftwareisrunning.TheActiveLEDoftheSCM(oroneS
6Troubleshooting123Blade and platform management issuesThetopicsinthissectionincludeShMC‐to‐IPMCcommunications,SELlocation,sensoreventrepo
6Troubleshooting124TypicallyeachIPMChasonlyitsowneventsinitsSEL.TheShMCsareexceptionstothisrulebecausetheylogeventsfromallIPMC
6Troubleshooting125Inventory Data Record (IDR) topicsRadisyscanprovideatoolcalledfruupdatethatcanbeusedtoupda teexistingFRUinventoryda
6Troubleshooting126Specific issuesTable 15listsspecificissuesthatcanbeencounteredwhileusingtheshelfmanagementandHPIservicesontheShel
6Troubleshooting127(continued from previous page)If the modules are still in debug mode, use these steps:• From each Shelf Manager module, verify that
6Troubleshooting128SCM resets abruptly forcing the LMP to reboot. There are no associated hot-swap events. Possible causes and steps to take are as fo
6Troubleshooting129Sessions cannot be opened on the HPI server. Error code: SA_ERR_HPI_OUT_OF_SPACEThe HPI server has reached the limit (32) for the n
1Shelf Management Overview13Figure 3. Shelf management component distribution in larger Radisys shelvesCarrier moduleProxy FRUIPMCOther FRUPEMIPMCIPMB
6Troubleshooting130RPT is incomplete; not all FRUs present in the shelf have associated Resources.1. Initiate a discovery operation using the CLI comm
6Troubleshooting131Resolving shelf FRU information issuesThissectionprovidesinformationonresolvingshelfFRUinformationproblems,includinghow
6Troubleshooting132Restoring shelf FRU informationTheprocedurebelowisusefulwhentheshelfFRUinformationisbadandneedstobereplacedinthe
6Troubleshooting133Restoring shelf FRU information on ATCA-6002 shelvesTheATCA‐6002shelfusesafiletostoreshelfFRUinformation,becausetherei
6Troubleshooting134IPMB addresses of slots and FRUsTable 16showstheIPMBaddressesofslotsintheATCA‐6002shelf.Table 17showstheIPMBaddresse
6Troubleshooting135Resource types of slots and FRUs in the RPTTheresourcepresencetableidentifiesthetypeofeachresourceinthe2ndbyteofthe
6Troubleshooting136Thefollowingaretheslotresourcetypesinhex:F0 UnknownslotF1 SPMorRCMslotF2 ShelfdisplaypanelslotF3 VirtualSPMorvirt
AAppendix137Shelf Manager InitializationShelf Manager and HPI interface initializationTheshelfmanagementserver(ShMS)isinstalledaspartofthel
AShelf Manager Initialization138d. PopulatestheRPTwithresourcesandtheirassociatedresourcedatarecords(RDRs).Resourcesarecreatedforeach
AShelf Manager Initialization139HPI subagent initializationWhentheHPIsubagentisstarted,initiallyyouwillseejustthestaticscalarobjectsdef
1Shelf Management Overview14Major elements of the shelf management infrastructureThefollowingsectionsdescribethemajorelementsoftheshelfmanag
BAppendix140Shelf Management FilesThisappendixdescribesfilesintheshelfmanagementinfrastructure,includingprograms,scripts,andconfiguration
BShelf Management Files141Shelf Manager log fileTheShelfManagerlogfileis/tmp/shmgr/shmgr.log.Backupfilesinthesamedirectoryarecrea tedas
CAppendix142IPMI Commands and Managed SensorsSupported IPMI commandsTheShelfManagersupportsallIPMIcommandslistedintheAdvancedTCABaseSpecif
CIPMI Commands and Managed Sensors143Table 18. Managed sensors for the Shelf Manager# Name Type CategoryNormal reading or event stateExplanation of
CIPMI Commands and Managed Sensors144Table 19. Managed sensors for each FRU# Name Type CategoryNormal reading or event stateExplanation of values256(0
CIPMI Commands and Managed Sensors14530 Get Address Info failed31 Self Test Error32 Self Test Check failed33 Set IPMB-<A|B> State <Enable|Dis
CIPMI Commands and Managed Sensors146258(0x102)FRU Thermal StatusTemperature Threshold 0x1 State 0x1: Lower minor.State 0x2: Lower major.State 0x4: L
1Shelf Management Overview15Field replaceable units (FRUs)AdvancedTCAdefinestwotypesofFRUsthatarevisibletoandcontrolledthroughtheIPMIin
1Shelf Management Overview16Management controllers (IPMCs and MMCs)EachintelligentFRUincludesamanagementcontroller,eitheranIntelligentPlatfo
1Shelf Management Overview17Theshelfcontainsanon‐volatilestoragedevicethatstoresshelfFRUinformation,includingdetailsoftheshelfitself
2Chapter18Software ArchitectureShelf management architectureShelfmanagementservicesfortheRadisysplatformarepartitionedacrosstwoprocessors:
2Software Architecture19Shelf interfaceAttheshelfinterfacelevel,theShMCfirmwarerunsontheIPMC,performingthelow‐levelshelfmanagementope
Revision historyVersion Date Description-0000 December 2006 First edition.-0001 September 2007 Second edition. New features and editorial corrections
2Software Architecture20HPI architectureTheprimarymanagementinterfacetotheRadisysshelfmanagementservicesistheHardwarePlatformInterface(
2Software Architecture21AsshowninFigure 5,theHPIclientlibrary(HCL)isadynamicallylinkedlibrary(DLL)thatprovidestheHPIAPIaswellas
2Software Architecture22RMCP,arequest‐responseprotocol,isasimplepacket‐basedcommunicationmechanism.RMCPpack etsincludeafieldthatindica
2Software Architecture23Forinformationonviewingthelogmessages,seeViewingthesystemeventlogonpage 101.Eventsmaybeclearedfromtheshel
2Software Architecture24Basic and enhanced shelf management operationTheShelfManagerinanATCA‐6002shelfcanoperateintwomodes:enhancedandba
2Software Architecture25Inenhancedmode,theShelfManagerusesthe/etc/shmgr.conffile.TheShelfManagerdoesnotneedtousetheshmgr_basic.conf
2Software Architecture26Figure 7. SCM Shelf Manager redundancy Shelf Manager failover ComponentfailuresintheShMSandShMCaredetectedthroughthe
2Software Architecture27Peer communication loss causes failoverIfcommunicationbetweenthepeerShMCsbreaksdown,thefirstofthepreviouscommunic
2Software Architecture28•UsingthefrontpanelresetbuttontoresettheactiveSCMeithermanuallyorusingHPIwiththesaHPIResourceResetStateSet()
2Software Architecture29Shelf FRU information redundancyDuringnormalshelfoperationwithredundantShelfManagers,therearealwaysatleastthreep
3Table of ContentsPreface ...
2Software Architecture30Shelf FRU invalid alarmThisalarmsignalsacatastrophicfaultcondition.TheactiveShelfManagerkeepsallothermodulesin
2Software Architecture31HPI server redundancyTheRadisy sHPIserverusesaredundancymodelverysimilartotheactive‐standbymodelusedbytheShel
2Software Architecture32InadditiontotheHPIdata,automaticoperationsaresynchronizedwiththepeerHPIserverstokeepHPIdatabasesinsync.Th
2Software Architecture33Shelf coolingToensurethatallmodulesintheshelfoperateefficiently,theiroperationaltemperaturesmustbemaintainedat
2Software Architecture34IfbothSCMsareremovedfromtheATCA‐6014orATCA‐6016shelf,theactiveRCMadjuststhefanstofullspeeduntiltheShelf
2Software Architecture35Shelf Manager cooling overrideTheShelfManagerprovidesautonomouscoolingmanagementfortheshelf.However,theremaybesi
3Chapter36Configuring Shelf Manager and HPI BehaviorThischapterexplainshowto:• DisableorenabletheRadisysShelfManager.• Changetheoperating
3Configuring Shelf Manager and HPI Behavior37Disabling the Shelf ManagerImportant:WhenthereisonlyoneSCMorShelfManagermoduleintheshelf,di
3Configuring Shelf Manager and HPI Behavior38Enabling the Shelf ManagerImportant:Ifyouareremovingamodulefromanon‐RadisysorRadisysMPCH0001
3Configuring Shelf Manager and HPI Behavior39Changing the Shelf Manager operating modeTheShelfManagerinanATCA‐6002shelfcanoperateintwomodes
4Table of ContentsChapter 4: Changing Shelf Settings Using HPI ... 53Radisys
3Configuring Shelf Manager and HPI Behavior40Overview of Shelf Manager and HPI configurationShelfManagerandHPIconfigurationsettingscanbeeither
3Configuring Shelf Manager and HPI Behavior41Table 4. Shelf Manager and HPI configuration settingsShelf Manager configuration settings Parameter to ed
3Configuring Shelf Manager and HPI Behavior42Editing the shmgr.conf configuration file File‐basedconfigurationsettingsarestoredinthe/etc/shmgr.
3Configuring Shelf Manager and HPI Behavior43Configuring Shelf Manager behaviorThissectiondescribestheshelfmanagementconfigurationsettingsyou
3Configuring Shelf Manager and HPI Behavior44Disabling the HPI serviceTheSHMS_CMDLINE_ARGSparameterin/etc/shmgr.confletsyoudisableHPIservice.
3Configuring Shelf Manager and HPI Behavior45Configuring the system event log (SEL)TheSHMS_SYNC_SEL_TIMEparameterin/etc/shmgr.confsetsthe syste
3Configuring Shelf Manager and HPI Behavior46Configuring HPI behaviorThissectiondescribestheHPIconfigurationsettingsyoucanchangebyeditingH
3Configuring Shelf Manager and HPI Behavior47• HSD_CHASSIS_FRU_MONITOR_ENABLEenablesthemonitoringofshelfFRUs,includingthealarmmoduleandfan
3Configuring Shelf Manager and HPI Behavior48Assigning an entity type for a new FRUTheHSD_FRU_ENTITY_TYPEparameterin/etc/shmgr.confletsyouassig
3Configuring Shelf Manager and HPI Behavior49Loading saved HPI resource configuration filesTheHSD_LOAD_SAVED_CONFIGURATIONparameterin/etc/shmgr.co
5Table of ContentsAppendix A: Shelf Manager Initialization... 137She
3Configuring Shelf Manager and HPI Behavior50Configuring the Shelf Manager in basic modeForaShelfManagerrunningontheATCA‐6002shelf,aconfigur
3Configuring Shelf Manager and HPI Behavior51Editing the rsyshsd.conf configuration fileForeachfrontmodulethereareblade HPIconfigurationsetti
3Configuring Shelf Manager and HPI Behavior52Configuring log settings•TheRSYSHSD_LOG_FILEparameteriswhereyoucanspecifythenameofthelogfile
4Chapter53Changing Shelf Settings Using HPIThischapterdescribestheRadisysHPIimplementation,includinghowHPIaffectsyourapplications,howto
4Changing Shelf Settings Using HPI54TheDomainIdneededforopeningHPIsessionsontheRadisysplatform’sHPIInterfaceisderivedfromtheShelfMan
4Changing Shelf Settings Using HPI55ThefollowingistheResourceIdformatusedbyshelfandShelfManagerresourcesintheHPIresourcepresencetabl
4Changing Shelf Settings Using HPI56Using the platform-management CLIYoucanaccessseveralcommandmodes,includingplatform‐managementmode,fromth
4Changing Shelf Settings Using HPI57Writing HPI applicationsThefollowingarerequiredtocreateanHPIapplication:SaHpi.h TheofficialHPIB.03.02h
4Changing Shelf Settings Using HPI58TorunanyHPIapplicationthatlinkswiththeRadisysHPIclientlibrary,libhcl.so,considerthefollowing:•The
4Changing Shelf Settings Using HPI59Using OpenHPI toolsOpenHPIisanopensourceprojectcreate dwiththeintentofprovidinganimplementationofSA
6PrefaceAbout this manualThismanualdescribestheRadisysShelfManagementSoftware(the“ShelfManager”),theHardwarePlatformInterface(HPI)serv
4Changing Shelf Settings Using HPI60Using SNMPSNMPaccesstotheshelfmanagementfunctionalityisprovidedthroughtheHPI‐B0101‐MIBmodule,whichwo
4Changing Shelf Settings Using HPI61Eachofthesetablesisindexedby:saHpiDomainId,saHpiResourceIdsaHpiResourceIsHistoricalsaHpiCtrl*EntryIdobject
4Changing Shelf Settings Using HPI62Configuring HPI subagent loggingHPIsubagentloggingisdisabledbydefault.Toenablelogging:1. Inthe/etc/rc.d
4Changing Shelf Settings Using HPI63Using the example HPI application (hpiapp)RadisysprovidesanexampleHPIapplication,hpiapp,thatiscompliantw
4Changing Shelf Settings Using HPI64Running hpiappThefollowingexamplesshowhowtorunthehpiappexampleapplicationtoconnecttoanHPIserverru
4Changing Shelf Settings Using HPI65Changing the shelf address Thefirst4bytesoftheshelfaddressshouldbechangedmanuallytouniquelyidentify
4Changing Shelf Settings Using HPI66Changing the Shelf Manager IP address TheactiveshelfmanagementserverhasadedicatedIPaddressthatisderive
4Changing Shelf Settings Using HPI67YoucaneitherchangeonelineatatimeorchangeallthreelinestogetherusingthesaHpiControlSet()function.
4Changing Shelf Settings Using HPI68Changing the FRU power-on sequence using the CLIYoucanusetheCLIto:•Viewthecurrentpower‐onsequence• Chang
4Changing Shelf Settings Using HPI69FRU power-on sequence controlsTheshelfresource(ResourceId0x01)providesasetofHPIdiscretecontrols,eacho
Preface7Shelves and modules supportedTheShelfManagerdescribedinthismanualisavailableontheseRadisysshelveswithasupportedfrontmodulein
4Changing Shelf Settings Using HPI70FRU power-on sequence commit status sensorTheshelfresourcecontainsanadditionalsensorcalledtheFRUpower‐on
4Changing Shelf Settings Using HPI71Querying the current hot-swap state of a FRUYoucanqueryaFRU’scurrenthot‐swapstate:•UsingtheCLIcommandst
4Changing Shelf Settings Using HPI72Overriding the default ShMS hot-swap mechanism WhetheraFRUsittinginaparticularslotisactivatedbytheShel
4Changing Shelf Settings Using HPI73Whileinmanualmode,youcanalsoupdatetheDelayBeforeNextPowerOnforthathardwareaddress.Thisisthean
4Changing Shelf Settings Using HPI74Overriding the cooling algorithmSystemmanagerscancompletelyoverridetheShelfManager’scoolingalgorithmand
4Changing Shelf Settings Using HPI75Overriding the cooling algorithm with an HPI applicationThefollowingstepsillustratehowyoucandothisusingH
4Changing Shelf Settings Using HPI76Changing shelf power propertiesTheshelfresourcecontainstwocontrolsthatenablesystemmanagerstoupdateshel
4Changing Shelf Settings Using HPI77HPI parameter controlTheshelfresources,alongwithallintelligentFRUresourcesandslotresources,supportthe
4Changing Shelf Settings Using HPI78HPI server parameter control configuration files Eachresourcehastwoconfigurationfiles:onefordefaultsettin
4Changing Shelf Settings Using HPI79How configuration files are created for a new resourceDuringHPIinitialization(whenresourcesarebeingcreated
Preface8Where to get more product informationVisittheRadisyswebsiteatwww.radisys.comforproductinformationandotherresources.Downloads(man
4Changing Shelf Settings Using HPI80Saving a configuration file to store custom settings for a resourceAfteraresourcehasbeenconfiguredtoadesir
4Changing Shelf Settings Using HPI81Error conditions TheparametercontroloperationsSAHPI_DEF A ULT_PARMorSAHPI_RESTORE_PARMwillfailandreturn
5Chapter82Managing Alarms and EventsThischapterdescribes:•TheShelfManager’sroleinmanagingalarms.•Howyoucanchangethealarmreportingbyass
5Managing Alarms and Events83Domain alarm types and troubleshooting overviewTable 8liststhealarmandsensortypesalongwithsampledescriptionsan
5Managing Alarms and Events84Threshold alarmsThefirsteightalarmtypesinTable 8arethresholdalarmswhichindicatethatathresholdsensordetect
5Managing Alarms and Events85UpperMajortemperatureeventsma yindicatethattheambienttemperatureistoohigh,inwhichca seyoumightneedtol
5Managing Alarms and Events86Resource utilization alarmsTheresourceutilizationalarmsindicatethattheShelfManagerisusingahighpercentageof
5Managing Alarms and Events87Domain alarm table entriesThealarmstringsarebasedontheassociatedsensorevents.Textwithinthealarmstringindic
5Managing Alarms and Events88Collecting information for the alarm configuration fileForeachtypeofFRUandeachsensorforwhichyouwill assignse
5Managing Alarms and Events89Table 11liststheinformationneededtoidentifythesensor,theeventstate,andthenewseveritytoassign.Table 11.
Preface9IPMIIntelligentPlatformManagementInterfaceSpecificationSecondGeneration,v2.0,DocumentRevision1.0,Intel,Hewlett‐Packard,NEC,Dell
5Managing Alarms and Events90Configuration file structure and contentsThe/etc/shmgralarm.conffileusesthisstructureforitscontents:FRUidentific
5Managing Alarms and Events91Alarm configuration procedureChangestothealarmconfigurationfiletakeeffectwhentheShelfManagerisrestarted.Thi
5Managing Alarms and Events92ATCA-6000 shelf display and alarm indicatorsTheshelfdisplaypanel(SDP)ontheATCA‐6000shelfconsistsofthefollowin
5Managing Alarms and Events93Figure 9. Format of alarm messages on the SDPSet to the location of the resource that has the alarm. The format is NNM fo
5Managing Alarms and Events94ATCA-6006 shelf alarm indicatorsTheoptionalshelfalarmpanelontheATCA‐6006shelfconsistsofthefollowing:•ShelfPo
5Managing Alarms and Events95ATCA-6014 and ATCA-6016 shelf alarm indicatorsTheATCA‐6014andtheATCA‐6016shelvesbothhaveashelfalarmpanel(SAP)
5Managing Alarms and Events96Thecontrolsforthetelcoalarmcontacts(AlarmCritical,AlarmMajor,andAlarmMinor)arestandardHPIcontrols.Onth
5Managing Alarms and Events97Parameter values for setting alarm controlsThefollowinglistidentifiesthepossiblesaHpiControlSet()functionparamete
6Chapter98TroubleshootingThischapterprovidesdetailedinformationonthefollowingissues:•AvailablediagnostictoolsforFRUandShelfManagerprob
6Troubleshooting99Using the Shelf Manager diagnostics utilityTheshmgrdiagutilitygathersinformationonthecurrentstateoftheShelfManager,the
Commentaires sur ces manuels