Sie sind auf Seite 1von 2

26/4/2015

StepstoreplacethesecondarypeeroftheDHCPFailoverassociationwiththenewmember

infoblox.com

WelcomeAlbertoHuamaniCanchizVILSOLSAC

SupportHome

KnowledgeBase

OpenaCase

Downloads

TechDocs

#223:Stepstoreplacethesecondarypeer
oftheDHCPFailoverassociationwiththe
newmember
Published11/01/201312:43PM|Updated11/01/201312:45PM

PROBLEMSUMMARY
DecommissioningandreplacinganexistingHApairmemberconfiguredasasecondarypeerof
theDHCPFOwiththenewHApairmember.

CUSTOMERENVIRONMENT
ANIOSGridrunningDHCPserviceswithDHCPFailover.

VERSION
AllNIOSversions.

CAUSE
WantedtoreplaceanexistingHAmemberconfiguredassecondarypeerofDHCPFailoverwith
thenewHAmember.

RESOLUTION
Detailsteps:
1. CreateanewdhcpfailoverassociationsandassignthenewHAmemberasa
secondarypeerwhereasprimarypeerwillbethesame.
2. UsingInfobloxAPIormanually,addthenewsecondarypeermemberIPaddresstothe
requirednetworks.
3. DisableDHCPserviceonoldsecondarypeer.
4. Settheprimarypeeroftheolddhcpfointhepartnerdownstate.Stepsareoutlinedat
theend.
5. UsingInfobloxAPI/manually,moveallrangestothenewdhcpfailoverassociations.
ClickonthetopRestartbuttontoRestarttheservices.
6. Makesurethehealthofthenewdhcpfoassociations.ItshouldshowNormalNormal
state.

SettingaPeerinthePARTNERDOWNState(primarypeer/survivingpeer)
Dothefollowing:
1. FromtheDataManagementtab,selecttheDHCPtab>Memberstab>Failover
Associations>failover_associationcheckbox.
2. ExpandtheToolbarandclickSetPartnerDown.
3. IntheSetFailoverAssociationPartnerDowndialogbox,selectoneofthefollowing:
Primary:Selectthisifthesecondaryserverisoutofservice.
Secondary:Selectthisiftheprimaryserverisoutofservice.
4. ClickOK.

WhatisPartnerDownState?
TheprimarypeerwillgointoastatecalledPartnerDownandwillanswerALLDHCPqueries.
DuringPartnerdownstate,theprimarypeerwillrenewtheclientleasetime=MCLT(bydefault
is3600secs).Formanyreasons,includinglongnetworkoutagesorplannedupgrades,one
peerofafailoverassociationmaybeinoperable.Inordertoprovideservicesduringthisoutage,
anadministratorcanputonepeerintoapartnerdownstate.Whenapeerisplacedinpartner
downstate,theremainingpeerassumescontrolofallDHCPservice.Itsafelyassumesthatits
peerisnotavailable,andcanstartassigningallIPaddressestoclientsneedingaddresses.As
https://support.infoblox.com/app/answers/detail/a_id/223/kw/change%20dhcp%20server

1/2

26/4/2015

StepstoreplacethesecondarypeeroftheDHCPFailoverassociationwiththenewmember

mentionedabove,anunexpectedfailurecouldcauseapeertogointothecommunications
interruptedstate,whichisnotadesirablelongtermstate.Ifapeergoesintoa
communicationsinterruptedstate,anadministratorcanforcetheremainingservertogointo
thertner
downstate.
Whenaserverchangesstate,itremembersthestarttimeofstate(STOS).Whenaserver
changestopartnerdown,itcanreclaimanyFREEorBACKUPaddressesthatbelongtoits
peerafterMCLTplusSTOShaspassed.IfanaddressisintheACTIVE,EXPIRED,
RELEASED,orRESETstate,andtheacknowledgedpotentialexpirytimeislaterthanSTOS,
theservercanfreetheIPaddressonlyaftertheacknowledgedpotentialexpirytimeplus
MCLT,orafterMCLTplusSTOS,whichevercomeslast.Thisisduetothefactthatthefailed
peermayhaveextendedtheleasetotheacknowledgedpotentialexpirytimeplusMCLT
withoutupdatingitspeer,butitmayhavealsoextendedittoMCLTplusSTOS,sothepeerin
thepartnerdownstatehastoaccountforbothpossibilities.
Afterapeerinthepartnerdownstatehasreclaimedalladdressesthatbelongtoitspeer,the
MCLThaspassed,andalltheacknowledgedpotentialexpirytimeshaveexpired,theremaining
peeroperatesasifitwerenolongerrunninginfailovermode.Itcontinuestodosountilitspeer
reconnectstoit.Duetothisfact,itisimperativethattherestartedpeernotprovideIPaddresses
untilithasresynchronizedwiththepeerinthepartnerdownstate.Thisprocedureisexplained
inthe"OperationsintheRECOVERState"section.Ifitwerenottodothis,thepotentialof
handingoutthewrongaddresstoclients,orduplicatingalreadyassignedaddresswouldbe
present.

ConsequencesWhenWeManuallyPutthePrimaryPeerinPartnerDown
1. Youwillnoticeanincreaseinnetworkbandwidthusageduetoleasetimeoffersequal
toMCLT(i.e.3600secs)
2. Therewillbeaminorperformanceissueontheprimarypeerboxdependingonthe
infobloxhardware.
NOTES:
Youcanperformstep1and2aheadoftimebeforeactuallyperformingmigration.
Makesurethatyouwilltakethedatabasebackupbeforeactuallyexecutestheabove
steps.
Afterperformingstep4,Westronglyrecommendyourteamtomovethecritical(larger)
networktothenewdhcpfoassociationandmonitorforsome
timeandcheckhowitgoes,forex,theclientgetsanipaddressorwhenclientrenews,
thedhcpserverrenewstheIPaddressetc.andthengoaheadandmovetherestofthe
ranges.
Afterassigningsomeoftherangestothenewdhcpfo,verifythestatusofthat,itshould
shownormalnormalstate.
Youdontneedtostopthedhcpserviceonthenewsecondarypeer.
Wecantsimplyreplacetheoldsecondarymemberwithnewsecondarymemberinthe
currentdhcpfo.NIOSwillthrowanerrorbecausetheold
secondarymemberismappedtotheranges/networksetc.
RecommendtocreateapreemptivecasewithInfobloxSupportbeforeperformingthe
migration.
Ifyouhaveanydoubts,pleasecontactInfobloxSupportpriortoperformingthisoperation.

https://support.infoblox.com/app/answers/detail/a_id/223/kw/change%20dhcp%20server

2/2

Das könnte Ihnen auch gefallen