• No results found

Replacing a faulty node with a spare node using the SAN Volume Controller Console

Toreplacea faultynode intheclusterusingtheSANVolumeControllerConsole, completethesetasks.

Before youattempttoreplaceafaultynodewithaspare nodeyoumustensure that:

v SANVolumeControllerversion 1.1.1orhigherisinstalledontheclusterandon thesparenode.Thecodelevels areavailablefromthenodeandclusterVPD.

Youcanalso issuethesvcinfo lsnodeorsvcinfo lsclustercommandsto determinetheSANVolumeControllerversion.

v Youknowthenameof theclusterthatcontainsthefaultynode.

v Asparenode isinstalledinthesamerack astheclusterthatcontainsthefaulty node.

v Youmakearecordof thelastfivecharactersof theoriginal worldwidenode name(WWNN)of thespare node.Youwillneedthisinformation, ifandwhen, youwanttostopusingthisnode asa sparenode. Inthatcase,youmightprefer touseit asanormalnodethatcanbeassignedtoanycluster.

PerformthefollowingstepstodisplayandrecordtheWWNNofthespare node:

1. Displaythenodestatusonthefrontpaneldisplayof thenode.SeetheIBM TotalStorageSANVolumeController:ServiceGuideformoreinformation.

2. Withthenodestatusdisplayedonthefrontpanel,pressandholdtheDown button;pressandreleasetheSelectbutton;releasetheDownbutton.WWNN isdisplayedonline-1of thezdisplay;line-2of thedisplaycontainsthelast fivecharactersoftheWWNN.

3. RecordtheWWNNina safeplace.Itwillbeneededif youwanttostopusing thesparenode.

If anodefails,theclustercontinuestooperatewithdegradedperformance,until the faultynodeisrepaired.If therepair operationislikelyto takeanunacceptable amountof time,it mightbeusefultoreplacethefaultynodewithasparenode.

However, theappropriateproceduresmust befollowedandprecautionsmust be taken,inordernottointerruptI/Ooperationsandtoavoidcompromisingthe integrity ofyourdata.Theproceduresoutlinedinthistopicinvolvechangingthe worldwidenodename (WWNN)of aSANVolumeController.Theseprocedures must befollowedwithcareinordertoavoidduplicateWWPNswhichcan cause data corruption.

Be awarethatbyperformingtheseproceduresthefollowingchangeswillbemade to yourconfiguration:

Front PanelID

This numberwillchange.It isthenumberthatisprintedonthefrontofthe node andusedtoselect thenode thatisto beaddedtoa cluster.

Node Name

This numbermightchange.Ifyou donotspecifyaname,theSANVolume Controllerassignsadefaultname whenaddinga nodetoa cluster.The SANVolumeControllercreatesanew nameeachtime anodeisaddedto a cluster.Ifyouchoose toassignyourown namesthenyouneedto typein

|

|

|

|

thenodename ontheAddinganode toaclusterpanel.Ifyou areusing scripts toperformmanagementtasksontheclusterandthosescriptsuse thenodename,thenbyassigningtheoriginalnameto areplacement node, youavoidtheneedto makechangestothescripts.

Node ID

This IDwillchange.Anewnode IDisassignedeach timeanodeisadded to acluster;thenodename remainsthesamefollowing serviceactivityon thecluster.Youcan usethenode IDorthenodenameto perform

management tasksonthecluster.However, ifyouareusingscriptsto performthosetasks, usethenode nameratherthanthenodeID.

Worldwide NodeName

This namewillchange.TheWWNNisused touniquelyidentifythenode andthefibre-channelports.TheWWNNof thesparenodewillchangeto thatof thefaultynode.Thenodereplacement proceduresmustbefollowed exactlytoavoidanyduplicationofWWNNs.

Worldwide PortNames

Thesenamesdonotchange.WWPNsarederivedfromtheWWNNthatis writtento thespare(replacement)node aspartof thisprocedure.For example,let’ssaytheWWNNforanode is50050768010000F6.Thefour WWPNsforthis nodewouldbederivedasfollows:

WWNN 50050768010000F6

WWNN displayed on front panel 000F6

WWPN Port 1 50050768014000F6

WWPN Port 2 50050768013000F6

WWPN Port 3 50050768011000F6

WWPN Port 4 50050768012000F6

Perform thefollowing stepsto replaceafaultynodeinthecluster:

1. VerifythenameandIDof thenode thatyouwishtoreplace.

PerformthefollowingstepstoverifythenameandID:

a. MakesurethattheSANVolumeControllerConsoleapplicationisrunning ontheclusterthatcontainsthefaultynode.

b. ClickWorkwithNodesintheportfolio.

c. ClickNodes.

Ifthenodeisfaulty,itwillbeshownasoffline.Ensure thepartnernodein theI/Ogroupisonline.

1) IftheothernodeintheI/Ogroup isoffline,startDirectedMaintenance Proceduresto determinethefault.

2) Ifyouhavebeen directedherebytheDMPs,andsubsequentlythe partnernodeintheI/Ogrouphasfailed,seetheprocedurefor RecoveringfromofflineVDisks.

Ifyouarereplacingthenode forotherreasons, determinethenodeyou wishtoreplaceandagainensurethepartner nodeintheI/Ogroup is online.

1) Ifthepartnernodeisoffline,youwillloseaccesstotheVDisksthat belongto thisI/Ogroupif youcontinue.StarttheDirectedMaintenance Proceduresandfixtheothernodebefore proceeding.

2. Findandrecordthefollowinginformationabout thefaultynode:

a. Nodename b. I/Ogroupname

c. Lastfivecharactersof theWWNN

d. FrontpanelID

e. Uninterruptiblepowersupplyserialnumber

a. Tofind andrecordthenodenameandI/Ogroup name,clickWorkwith Nodesintheportfolio.

b. ClickNodes.

Thefaultynodewillbeoffline.

c. Recordthefollowinginformationaboutthefaultynode:

v Nodename v I/Ogroupname

d. Tofind andrecordthelastfivecharactersof theWWNN,clickonthename oftheofflinenode.

e. ClicktheGeneraltab.

f. RecordthelastfivecharactersoftheWWNN.

g. Tofind andrecordthefrontpanelID,clicktheVital ProductDatatab.

h. Findthefront-panel-assemblysection ofthevitalproductdata (VPD).

i. RecordthefrontpanelID.

j. Tofind andrecordtheuninterruptiblepowersupplyserialnumber,clickthe VitalProductDatatab.

k. Findtheuninterruptiblepowersupply sectionoftheVPD.

l. Recordtheuninterruptiblepowersupplyserialnumber.

3. ObtaintheIDofthefaultynode.Disconnectallfourfibre-channelcablesfrom thenode.

Important:Donotplugthefibre-channelcablesintothespare nodeuntil sparenodehas beenconfiguredwiththeWWNNfromthefaultynode.

4. Connectthepowerandsignalcablesfromthesparenode tothe

uninterruptiblepowersupply thathas theserialnumberthatyounotedinstep 5l.

Note: Thesignalcablecanbepluggedintoanyvacantpositiononthetoprow of serialconnectors ontheuninterruptiblepowersupply.Ifnospare serialconnectorsareavailableontheuninterruptiblepowersupply, disconnectthecablesfromthefaultySANVolumeController.

5. Power-onthesparenode.

6. Displaythenodestatusontheservice panel.SeetheIBMTotalStorage SAN VolumeController:ServiceGuideformoreinformation.

7. ChangetheWWNNofthesparenode.

PerformthefollowingstepstochangetheWWNNofthespare nodesothatit matchestheWWNNofthefaultynode:

a. Withthenode statusdisplayedonthefrontpanel,pressandhold the Downbutton;pressandreleasetheSelectbutton;releasetheDown button.WWNNisdisplayedonline-1ofthedisplay;line-2of thedisplay containsthelastfive charactersoftheWWNN.

b. WiththeWWNNdisplayedontheservicepanel,pressandholdtheDown button,pressandreleasetheSelectbutton,releasetheDownbutton.This switchesthedisplayintoedit mode.Change thedisplayed numberto matchtheWWNNrecordedinstep5f.

Note: ToeditthedisplayednumberusetheUpandDownbuttonsto increaseordecreasethenumbers displayed.Usetheleft andright

buttonsto movebetweenfields.Whenthefivecharactersmatchthe numberrecordedinstep1,presstheselectbuttontwiceto accept thenumber.

8. Connectthefourfibre-channelcables thatweredisconnectedfromthefaulty nodetothespare node.

9. RemovethefaultynodefromtheclusterusingtheSANVolumeController Console.

Remember:Recordthefollowinginformation:

v Nodeserialnumber v WWNN

v AllWWPNs

v I/Ogroupthatcontainsthenode

Thiscanavoidapossibledatacorruptionexposurewhenthenodeisre-added tothecluster.

10. Addthespare nodetotheclusterusingtheSANVolumeControllerConsole.

11. UsetheSubsystemDeviceDriver(SDD)managementtool onthehost systemsto verifythatallpathsarenowonline.SeetheIBMTotalStorage SubsystemDeviceDriver:User’s Guideformoreinformation.

Attention: Whenthefaultynode isrepaireddonotconnectthefibre-channel cablestoit.Connectingthecables mightcausedatacorruption.

12. Repairthefaultynode.

13. Ifyouwanttousetherepairednodeasaspare node,performthefollowing steps:

a. Displaythenode statusonthefrontpaneldisplayof thenode. Seethe IBMTotalStorage SANVolumeController:ServiceGuidefor more information.

b. Withthenode statusdisplayedonthefrontpanel,pressandhold the Downbutton;pressandreleasetheSelect button;releasetheDown button.WWNNisdisplayedonline-1of thedisplay;line-2of thedisplay containsthelastfive charactersoftheWWNN.

c. WiththeWWNNdisplayed ontheservicepanel,pressandholdtheDown button,pressandreleasetheSelectbutton,releasetheDownbutton.This switchesthedisplayintoeditmode.Change thedisplayed numberto 00000.

Note: ToeditthedisplayednumberusetheUpandDownbuttons to increaseordecreasethenumbersdisplayed.Usetheleft andright buttonsto movebetweenfields.

ThisSANVolumeControllercannowbeusedasaspare node.

Attention: Neverconnecta SANVolumeControllerwithaWWNNof00000 tothecluster.If thisSANVolumeControllerisnolongerrequiredasaspare andisto beusedfornormalattachmentto aclusteryoumustfirst usethe proceduredescribedinthe″Prerequistes″ tochangetheWWNNtothenumber yourecordedwhenasparewascreated.Usinganyothernumbermightcause datacorruption.

Relatedtasks

“Deletinganodefromacluster”onpage155

Anodemightneedtobedeletedfromaclusterif thenodehasfailedandis beingreplaced witha newnodeor iftherepairthathasbeenperformed has causedthatnodeto beunrecognizable bythecluster.

“LaunchingtheSANVolumeControllerConsole”onpage114

YoucanlaunchtheSANVolumeControllerfromtheViewingClusterspanel.

“Replacinga faultynode intheclusterusingtheCLI”onpage 206

Youcanreplaceafaultynodeintheclusterusingthecommand-lineinterface (CLI).

Relatedinformation

“Advancedfunctionsoverviewfor theSANVolumeControllerConsole”onpage 137

Youcanperform″advanced″functionsusingtheSANVolumeController Console.