RAC/ASM/VOTING DISK Interview Questions & Answer
Q What is SCAN?
Single Client Access Name (SCAN) is s a new Oracle Real Application Clusters (RAC) 11g Release 2 feature that provides a single name na me for clients to access an Oracle Dataase running in a cluster! "he enefit is clients using SCAN do not need to change if #ou add or remove nodes in the cluster! Q what is !na"i# re"asterin$ ? When wi%% the !na"i# re"asterin$ haens?
d#namic remastering is ailit# to move the ownership of resource from one instance to another instance in RAC! d#namic resource remastering is used to implement for resource affinit# for increased performance! resource affinit# optimi$ed the s#stem in situation where update transactions are eing e%ecuted in one instance! when activit# shift to another instance the resource affinit# correspondingl# move to another instance! &f activit# is not locali$ed then resource ownership is hashed to the instance! &n 1'g d#namic remastering happens in fileoect level!the process of remastering is ver# stringent! *or one instance should touch more than +' times than the other instance in particular period(sa# 1' mints)! this touch ratio and time can e tuned # gc,affinit#,limit gc,affinit#,limit and ,gc,affinit#,time parameter! parameter! Q wh! we re'uire to "aintain o nu"(er o) votin$ is*s?
Odd numer of dis- are to avoid a void split rain. /hen Nodes in cluster can0t tal- to each other the# run to loc- the oting oting dis- and whoever loc- the more dis- will survive. if dis- numer are even there are chances that node might loc- +' of dis- (2 out of 3) then how to decide which node to evict! whereas when numer is odd. one will e higher than other and each for cluster to evict the node with less numer Q +ow !ou #he#* the hea%th o) ,our RAC Data(ase? -#rs#t%- command from root or oracle user can e used to chec- the clusterware health 4ut for starting or stopping we have to use root user or an# privilege user!
5oracle6"7S",NOD71 89: #rs#t% #he#* #rs CSS appears health#
CRS appears health# 7; appears health# Q +ow !ou #he#* the servi#es in RAC Noe?
/e can chec- the service or start the services with 0srvctl0 command!load alanced<"A* service named RAC online! 5oracle6"7S",NOD71 89: srvctl start service =d orcl =s RAC 5oracle6"7S",NOD71 89: crsstat Q I) there is so"e issue with virtua% I. how wi%% !ou trou(%eshoot it?+ow wi%% !ou #han$e virtua% i?
"o change the &> (virtual (virtual &>) on a RAC node. use the command 5oracle6testnode oracle9: srv#t% "oi)! nodeapps =A new,address Q +ow !ou wi%% (a#*u !our RAC Data(ase?
4ac-up strateg# of RAC Dataase? An RAC Dataase consists of 0OCR 10Votin$ 10Votin$ is* & 20Data(ase )i%es3 #ontro%)i%es3 reo%o$ )i%es & Ar#hive %o$ )i%es Q Do !ou have an! iea o) %oa (a%an#in$ in a%i#ation?+ow %oa (a%an#in$ is one?
http?<
RAC stands for Real Application cluster! &t is a clustering solution from Oracle Corporation that ensures high availailit# of dataases # providing instance failover. media failover features! Q What is RAC an how is it i))erent )ro" non RAC ata(ases?
RAC stands for Real Application Cluster. Cluster. #ou have n numer of o f instances running in their own separate nodes and ased on the shared storage! Cluster is the -e# component and is a collection of servers operations as one unit! RAC is the est est solution for high performance and high availal#! Non RAC dataases has single point of failure in case of hardware failure or server crash! Q Give the usa$e o) srv#t% ?
srvctl start instance =d d,name =i @inst,name,list@ 5=o start,options9 srvctl stop instance =d name =i @inst,name,list@ 5=o stop,options9 srvctl stop instance =d orcl =i @orcl.orcl3@ =o immediate srvctl start dataase =d name 5=o start,options9 srvctl stop dataase =d name 5=o stop,options9 srvctl start dataase =d orcl =o mount Q Mention the Ora#%e RAC so)tware #o"onents ?
Oracle RAC is composed of two or more dataase instances! "he# are composed of ;emor# structures and ac-ground processes same as the single instance dataase!Oracle RAC instances use two processes B7S(Bloal 7nueue Service). BCS(Bloal Cache Service) that enale cache fusion!Oracle RAC instances are composed of following ac-ground processes? AC;SAtomic Controlfile to ;emor# Service (AC;S) B"E'=Bloal "ransaction >rocess F;ONBloal 7nueue Service ;onitor F;DBloal 7nueue Service Daemon F;SBloal Cache Service >rocess FCG'&nstance 7nueue >rocess R;SnOracle RAC ;anagement >rocesses (R;Sn) RS;NRemote Slave ;onitor Q What is GRD?
BRD stands for Bloal Resource Director#! "he B7S and BCS maintains records of the statuses of each datafile and each cahed loc- using gloal resource director#!"his process is referred to as cache fusion and helps in data integrit#! Q What are the i))erent networ* #o"onents are in 4$ RAC?
pulic. private. and vip components >rivate interfaces is for intra node communication! &> is all aout availailit# of application! /hen a node fails then the &> component fail over to some other node. this is the reason that all applications should ased on vip components means tns entries should have vip entr# in the host list Q Give Detai%s on ACMS5
AC;S stands for Atomic Controlfile ;emor# Service!&n an Oracle RAC environment AC;S is
an agent that ensures a distriuted SBA memor# update(ie)SBA updates are gloall# committed on success or gloall# aorted in event of a failure! Q What are the "a6or RAC wait events?
&n a RAC environment the uffer cache is gloal across all instances in the cluster and hence the processing differs!"he most common wait events related to this are gc cr reuest and gc uffer us# BC CR reuest ?the time it ta-es to retrieve the data from the remote cache Reason? RAC "raffic Hsing Slow Connection or &nefficient ueries (poorl# tuned ueries will increase the amount of data loc-s reuested # an Oracle session! "he more loc-s reuested t#picall# means the more often a loc- will need to e read from a remote instance via the interconnect!) BC 4H**7R 4HSI? &t is the time the remote instance locall# spends accessing the reuested data loc-! Q Give etai%s on GT7486
"he process provides transparent support for EA gloal transactions in a RAC environment!"he dataase autotunes the numer of these processes ased on the wor-load of EA gloal transactions! Q Give etai%s on 9MON
"his process monitors gloal enues and resources across the cluster and performs gloal enueue recover# operations!"his is called as Bloal 7nueue Service ;onitor! Q Give etai%s on 9MD
"his process is called as gloal enueue service daemon! "his process manages incoming remote resource reuests within each instance! Q Give etai%s on 9MS
"his process is called as Bloal Cache service process!"his process maintains statuses of datafiles and each cahed loc- # recording information in a Bloal Resource Dector#(BRD)!"his process also controls the flow of messages to remote instances and manages gloal data loc- access and transmits loc- images etween the uffer caches of different instances!"his processing is a part of cache fusion feature! Q Give etai%s on 9CK4
"his process is called as &nstance enueue process!"his process manages non=cache fusion resource reuests such as lir# and row cache reuests! Q Give etai%s on RMSn
"his process is called as Oracle RAC management process!"hese pocesses perform managailit# tas-s for Oracle RAC!"as-s include creation of resources related Oracle RAC when new instances are added to the cluster!
Q +ow to e:ort an i"ort #rs resour#es whi%e "i$ratin$ Ora#%e RAC to new server;
4elow script generate svrctl add script for dataase. instance. service a nd 11B listeners from OCR from current RAC! Save the result of the script and run it at new RAC! for D4NA;7 in :(srvctl config dataase) do J Benerate D4 resource srvctl config dataase =d :D4NA;7 =a K aw- =v dnameL@:D4NA;7@ M 047B&N *SL@?@ :18
*&F7< spfile L @=p@ :2 :18
:28<2:< printf @s s s s s Mn@. @srvctl modif# instance =d @.dname. @ =i @. :2 .@ =s AS;2@ :28<:< printf @s s s s s Mn@. @srvctl modif# instance =d @.dname. @ =i @. :2 .@ =s AS;@ :28<3:< printf @s s s s s Mn@. @srvctl modif# instance =d @.dname. @ =i @. :2 .@ =s AS;3@ 0 fi echo @srvctl start dataase =d :D4NA;7@ J Benerate Service resource snamelistL:(srvctl status service =d :D4NA;7 K aw- 0print :20) for sname in :snamelist do srvctl config service =d :D4NA;7 =s :snameK aw- =v dnameL@:D4NA;7@ =v snameL:sname M 047B&N *SL@?@ :18<>referred instances< pref L @=r@ :2 :18<>R7*< pref L @=r@ :2Q su( 4AS&C@0 echo @srvctl start service =d :D4NA;7 =s :sname@ done done J Fistener at 11B Pome! 1'B listener can0t a added with srvctl!
srvctl config listener K aw- M 047B&N *SL@?@Q state L 'Q :18
Q Give etai%s on RSMN
"his process is called as Remote Slave ;onitor!"his process manages ac-ground slave process creation andd communication on remote instances! "his is a ac-ground slave process!"his process performs tas-s on ehalf of a co=ordinating process running in another instance! Q What #o"onents in RAC "ust resie in share stora$e?
All datafiles. controlfiles. S>*&les. redo log files must reside on cluster=aware shred storage! Q What is the si$ni)i#an#e o) usin$ #%uster8aware share stora$e in an Ora#%e RAC environ"ent?
All instances of an Oracle RAC can access all the datafiles.control files. S>*&F70s. redolog files when these files are hosted out of cluster=aware shared storage which are group of shared dis-s! Q Give )ew e:a"%es )or so%utions that suort #%uster stora$e
AS;(automatic storage management).raw dis- devices.networ- file s#stem(N*S). OC*S2 and OC*S(Oracle Cluster *ie s#stems)! Q What is an inter#onne#t networ*?
An interconnect networ- is a private networ- that connects all of the servers in a cluster! "he interconnect networ- uses a switch
Configure Hser Datagram >rotocol(HD>) on Bigait ethernet for cluster interconnect!On uni% and linu% s#stems we use HD> and RDS(Reliale data soc-et) protocols to e used # Oracle Clusterware!/indows clusters use the "C> protocol! Q Can we use #rossover #a(%es with Ora#%e C%usterware inter#onne#ts?
No. crossover cales are not supported with Oracle Clusterware intercnects! Q What is the use o) #%uster inter#onne#t?
Cluster interconnect is used # the Cache fusion for inter instance communication! Q +ow o users #onne#t to ata(ase in an Ora#%e RAC environ"ent?
Hsers can access a RAC dataase using a client
to dataase! Q What is the use o) a servi#e in Ora#%e RAC environ"ent?
Applications should use the services feature to connect to the Oracle dataase!Services enale us to define rules and characteristics to control how users and applications connect to dataase instances! Q What are the #hara#teristi#s #ontro%%e (! Ora#%e servi#es )eature?
"he charateristics include a uniue name. wor-load alancing and failover options.and high availailit# characteristics! Q What ena(%es the %oa (a%an#in$ o) a%i#ations in RAC?
Oracle Net Services enale the load alancing of application connections across all of the instances in an Oracle RAC dataase! Q What is a virtua% I. aress or VI.?
A virtl &> address or &> is an alternate &> address that the client con nectins use instead of the standard pulic &> address! "o configure&> address. we need to reserve a spare &> address for each node. and the &> addresses must use the same sunet as the pulic networ-! Q What is the use o) VI.?
&f a node fails. then the node0s &> address fails over to another node on which the &> address can accept "C> connections ut it cannot accept Oracle connections! Q Give situations uner whi#h VI. aress )ai%over haens
&> addresses failover happens when the node on which the &> address runs fails. all interfaces for the &> address fails. all interfaces for the &> address are disconnected from the networ-! Q What is the si$ni)i#an#e o) VI. aress )ai%over?
/hen a &> address failover happens. Clients that attempt to connect to the &> address receive a rapid connection refused error !"he# don0t have to wait for "C> connection timeout messages! Q What are the a"inistrative too%s use )or Ora#%e RAC environ"ents?
Oracle RAC cluster can e administered as a single image using O7;(7nterprise ;anager).SVFW>FHS.Servercontrol(SRC"F).clusterverificationutilit#(cvu).D4CA.N7"CA
Q +ow o we veri)! that RAC instan#es are runnin$?
&ssue the following uer# from an# one node connecting through SVFW>FHS! :connect s#s
*ast application Notification as it areviates to *AN relates to the events related to instances.services and nodes!"his is a notification mechanism that Oracle RAc uses to notif# other processes aout the configuration and service level information that includes service status changes such as.H> or DO/N events!Applications can respond to *AN events and ta-e immediate action! Q Where #an we a%!
*AN H> and *AN DO/N events can e applied to instances.services and nodes! State the use of *AN events in case of a cluster configuration changeY During times of cluster configuration changes.Oracle RAC high availailit# framewor pulishes a *AN event immediatel# when a state change occurs in the cluster!So applications can receive *AN events and react immediatel#!"his prevents applications from polling dataase and detecting a prolem after such a state change! Q Wh! shou% we have seerate ho"es )or AS" instan#e?
&t is a good practice to have AS; home seperate from the dataase hom(ORACF7,PO;7)!"his helps in upgrading and patching AS; and the Oracle dataase software independent of each other!Also.we can deinstall the Oracle dataase software independent of the AS; instance! Q What is the avanta$e o) usin$ ASM?
Paving AS; is the Oracle recommended storage option for RAC dataases as the AS; ma%imi$es performance # managing the storage configuration across the dis-s!AS; does this # distriuting the dataase file across all of the availale storage within our cluster dataase environment! Q What is ro%%in$ u$rae?
&t is a new AS; feature from Dataase 11g!AS; instances in Oracle dataase 11g release(from 11!1) can e upgraded or patched using rolling upgrade feature! "his enales us to patch or upgrade AS; nodes in a clustered environment without affecting dataase availailit#!During a rolling upgrade we can maintain a functional cluster while one or more of the nodes in the
cluster are running in different software versions! Q Can ro%%in$ u$rae (e use to u$rae )ro" 4$ to $ ata(ase?
No.it can e used onl# for Oracle dataase 11g releases(from 11!1)! Q State the initia%i>ation ara"eters that "ust have sa"e va%ue )or ever! instan#e in an Ora#%e RAC ata(ase
Some initiali$ation parameters are critical at the dataase creation time and must have same values!"heir value must e specified in S>*&F7 or >*&F7 for ever# instance!"he list of parameters that must e identical on ever# instance are given elow? AC"&7,&NS"ANC7,COHN" ARCP&7,FAB,"ARB7" CO;>A"&4F7 CFHS"7R,DA"A4AS7 CFHS"7R,DA"A4AS7,&NS"ANC7 CON"ROF,*&F7S D4,4FOCG,S&Z7 D4,DO;A&N D4,*&F7S D4,NA;7 D4,R7CO7RI,*&F7,D7S" D4,R7CO7RI,*&F7,D7S",S&Z7 D4,HN&VH7,NA;7 &NS"ANC7,"I>7 (RD4;S or AS;) >ARAFF7F,;AE,S7R7RS R7;O"7,FOB&N,pass/ORD,*&F7 HNDO,;ANAB7;7N" Q What is ORA844425 ORAC9@ server session ter"inate (! )ata% error or ORA81B415 error o##urre in C%uster Grou Servi#e oeration?
RAC node name was listed in the loopac- address!!! Q Can the DM99OCKS an R@S=9TCAC+@MA7SI@ (e ienti#a% on a%% instan#es?
"hese parameters can e identical on all instances onl# if these parameter values are set to $ero! /hat two parameters must e set at the time of starting up an AS; instance in a RAC environmentY"he parameters CFHS"7R,DA"A4AS7 and &NS"ANC7,"I>7 must e set! Q Mention the #o"onents o) Ora#%e #%usterware
Oracle clusterware is made up of components li-e voting dis- and Oracle Cluster
Registr#(OCR)! Q What is a CRS resour#e?
Oracle clusterware is used to manage high=availailit# operations in a cluster!An#thing that Oracle Clusterware manages is -nown as a CRS resource!Some e%amples of CRS resources are dataase.an instance.a service.a listener.a &> address.an application process etc! Q What is the use o) OCR?
Oracle clusterware manages CRS resources ased on the configuration information of CRS resources stored in OCR(Oracle Cluster Registr#)! Q +ow oes a Ora#%e C%usterware "ana$e CRS resour#es?
Oracle clusterware manages CRS resources ased on the configuration information of CRS resources stored in OCR(Oracle Cluster Registr#)! Q Na"e so"e Ora#%e #%usterware too%s an their uses?
O&*C*B = allocating and deallocating networ- interfaces OCRCON*&B = Command=line tool for managing Oracle Cluster Registr# OCRDH;> = &dentif# the interconnect eing used CH = Cluster verification utilit# to get status of CRS resources Q What are the "oes o) e%etin$ instan#es )ro" ORa#%e Rea% A%i#ation #%uster Data(ases?
/e can delete instances using silent mode or interactive mode using D4CA(Dataase Configuration Assistant)! Q +ow o we re"ove ASM )ro" a Ora#%e RAC environ"ent?
/e need to stop and delete the instance in the node first in interactive or silent mode!After that asm can e removed using srvctl tool as follows? srvctl stop asm =n node,name srvctl remove asm =n node,name /e can verif# if AS; has een removed # issuing the following command? srvctl config asm =n node,name Q +ow o we veri)! that an instan#e has (een re"ove )ro" OCR a)ter e%etin$ an instan#e?
&ssue the following srvctl command? srvctl config dataase =d dataase,name
cd CRS,PO;7<in !
/e can verif# the current ac-up of OCR using the following command ? ocrconfig =showac-up /hat are the performance views in an Oracle RAC environmentY /e have v: views that are instance specific! &n addition we have B: views called as gloal views that has an &NS",&D column of numeric data t#pe!B: views otain information from individual : views! /hat are the t#pes of connection load=alancingY "here are two t#pes of connection load=alancing?server=side load alancing and client=side load alancing! Q What is the i))eren#e (etween server8sie an #%ient8sie #onne#tion %oa (a%an#in$?
Client=side alancing happens at client side where load alancing is done using listener!&n case of server=side load alancing listener uses a load=alancing advisor# to redirect connec tions to the instance providing est service! Q What are the three $reatest (ene)its that RAC rovies??
"he three main enefits are availailit#. scalailit#. and the ailit# to use low cost commodit# hardware! RAC allows an application to scale verticall#. # adding C>H. dis- and memor# resources to an individual server! 4ut RAC also provides hori$ontal scalailit#. which is achieved # adding new nodes into the cluster! RAC also allows an organi$ation to ring these resources online as the# are needed! "his can save a small or midsi$e organi$ation a lot of mone# in the earl# stages of a proect! &n a RAC environment. if a node in the cluster fails. the application continues to run on the surviving nodes contained in the cluster! &f #our application is configured correctl#. most users won0t even -now that the node the# were running on ecame unavailale! Q What are the "a6or RAC wait events?
&n a RAC environment the uffer cache is gloal across all instances in the cluster and hence the processing differs!"he most common wait events related to this are $# #r re'uest and gc uffer us# GC CR re'uest ? the time it ta-es to retrieve the data from the remote cache
Reason? RAC "raffic Hsing Slow Connection or &nefficient ueries (poorl# tuned ueries will increase the amount of data loc-s reuested # an Oracle session! "he more loc-s reuested t#picall# means the more often a
loc- will need to e read from a remote instance via the interconnect!) GC E=<<@R E=S, ? &t is the time the remote instance locall# spends accessing the reuested data loc-! Q What are the i))erent networ* #o"onents in Ora#%e 4$ RAC?
/e have pulic. private. and &> components! >rivate interfaces is for intra node communication! &> is all aout availailit# of application! /hen a node fails then the &> component will fail over to some other node. this is the reason that all applications should e ased on &> components! "his means that tns entries should have &> entr# in the host list! Q Tune the )o%%owin$ RAC DATAEAS@ FDENAM@.ROD0 whi#h is 2 noe RAC;
>ROD1 C>H 2 B4 RA;
>ROD2 C>H 1+ 12 B4 RA;
>ROD C>H 1[ B4 RA;
/hat are #ou loo-ing for hereY /hat tuning information do #ou e%pectY &t is a node cluster with different hardware configuration running RAC! & would put 2' of the memor# for Oracle in each node! So that would mean that the SBA is different in each of the nodes! Also since the C>H0s are different >ROD2 can have more numer of ma% numer of processes as compared to the rest of them! 4ut as & said this is ust configuration. this is not tuning! Vuestion is not clear! Q Write a sa"%e s#rit )or RMAN )or the re#over! i) a%% the instan#e are own;F
4ring all nodes down! Start one Node Restore all datafiles and archive logs! Recover 1 Node! Open the dataase! ring other nodes up! Confirm that all nodes are operational! C%ients are er)or"in$ so"e oeration an suen%! one o) the ata)i%e is e:erien#in$ ro(%e" what o !ou o? The #%uster is a two noe one;
1! 4ring the datafile offline recover the datafile!
1! +ow #an !ou #onne#t to a se#i)i# noe in a RAC environ"ent? 2! tnsnames!ora ensure that #ou have &NS"ANC7,NA;7 specified in it! Q +ow to "ove OCR an Votin$ is* to new stora$e evi#e? Movin$ OCR Iou must e logged in as the root user. ecause root owns the OCR files! Also an ocrmirror must e in place efore tr#ing to replace the OCR device!
;a-e sure there is a recent ac-up of the OCR file efore ma-ing an# changes? ocrconfig –showbackup
&f there is not a recent ac-up cop# of the OCR file. an e%port can e ta-en for the current OCR file! Hse the following command to generate an e%port of the online OCR file? &n 1'!2 J ocrconfig –export -s online &n 11g J ocrconfig -manualbackup "he new OCR dis- must e owned # root. must e in the oinstall group. and must have permissions set to [3'! >rovide at least 1'' ;4 dis- space for the OCR! On one node as root run? # ocrconfig -replace ocr # ocrconfig -replace ocrmirror
Now run ocrchec- to verif# if the OCR is pointing to the new file Movin$ Votin$ Dis*
Note? crsctl votedis- commands must e run as root Shutdown the Oracle Clusterware (crsctl stop crs as root) on all nodes efore ma-ing an# modification to the voting dis-! Determine the current voting dis- location using? crsctl query css votedisk
"a-e a ac-up of all voting dis-? dd if=voting_disk_name of=backup_file_name
"o move a oting Dis-. provide the full path including file name? crsctl delete css votedisk –force crsctl add css votedisk –force
After modif#ing the voting dis-. start the Oracle Clusterware stac- on all no des # crsctl start crs
erif# the voting dis- location using crsctl query css votedisk Q What is run)i:u;sh s#rit in Ora#%e C%usterware $ re%ease 1 insta%%ation
/ith Oracle Clusterware 11g release 2. Oracle Hniversal &nstaller (OH&) detects wh en the minimum reuirements for an installation are not met. and creates shell scripts. called fi%up scripts. to finish incomplete s#stem configuration steps! &f OH& detects an incomplete tas-. then it generates fi%up scripts (runfi%up!sh)! Iou can run the fi%up script after #ou clic- the *i% and Chec- Again 4utton! "he *i%up script does the following? \ &f necessar# sets -ernel parameters to values reuired for successful installation. including? ] Shared memor# parameters! ] Open file descriptor and HD> send
\ Sets shell limits if necessar# to reuired values! Q When e:a#t%! urin$ the insta%%ation ro#ess are #%usterware #o"onents #reate?
After fulfilling the pre=installation reuirements. the asic installation steps to follow are? ! &nvo-e the Oracle Hniversal &nstaller (OH&) 1! 7nter the different information for some components li-e? = name of the cluster = pulic and private node names = location for OCR and oting Dis-s = networ- interfaces used for RAC instances =etc! 2! After the Summar# screen. OH& will start cop#ing under the :CRS,PO;7 (this is the :ORACF7,PO;7 for Oracle Clusterware) in the local node the liraries and e%ecutales! = here we will have the daemons and scripts init!W created and configured properl#!
Oracle lusterware is formed of several daemons! each one of which ha ve a special function inside the stack" aemons are executed via the init"$ scripts %init"cssd! init"crsd and init"evmd&"
= note that for CRS onl# some client liraries are recreated. ut not all the e%ecutales (as for the RD4;S)! H! Fater the software is propagated to the rest of the nodes in the cluster and the ora&nventor# is updated! ! "he installer will as- to e%ecute root!sh on each node! Hntil this step the software for Oracle Clusterware is inside the :CRS,PO;7!
Running root!sh will create several components outside the :CRS ,PO;7? = OCR and D will e formated! = control files (or SCFS,SRC files ) will e created with the correct co ntents to start Oracle Clusterware! 'hese files are used to control some aspects of Oracle lusterware like( - enable)disable processes from the ** family %+g" oprocd! oslsvmon& - stop the daemons %ocssd"bin! crsd"bin! etc&" - prevent Oracle lusterware from being started when the machine boots" - etc"
=
= the different processes init!W (init!cssd. init!crsd. etc) will start the daemons (ocssd!in. crsd!in. etc)! /hen all the daemons are running then we can sa# that the installation was successful = On 1'!2 and later. running root!sh on the last node in the cluster also will create the nodeapps (&>. BSD and ONS)! On 1'!1. &>CA is e%ecuted as part of the RAC installation! ! After running root!sh on each node. we need to continue with the OH& session! After pressing the 0OG0 utton OH& will include the information for the pulic and cluster,interconnect interfaces! Also CH (Cluster erification Htilit#) will e e%ecuted! Q What are Ora#%e C%usterware ro#esses )or 4$ on =ni: an 9inu:
luster *ynchroniation *ervices %ocssd& . ;anages cluster node memership and runs as the oracle userQ failure of this process results in cluster restart! luster /eady *ervices %crsd& "he crs process manages cluster resources (which could e a dataase. an instance. a service. a Fistener. a virtual &> (&>) address. an application process. and so on) ased on the resource0s configuration information that is stored in the OCR! "his includes start. stop. monitor and failover operations! "his process runs as the root user +vent manager daemon %evmd& . A ac-ground process that pulishes events that crs creates! 0rocess 1onitor aemon %O0/O& . "his process monitor the cluster and provide &ROCD performs its chec-. stops running. and if the wa-e up is e#ond the e%pected time. then O>ROCD resets the processor and reoots the node! An O>ROCD failure results in Oracle Clusterware restarting the node! O>ROCD uses the hangchec- timer on Finu% platforms! /23 %racgmain! racgimon& 7%tends clusterware to support Oracle=specific reuirements and comple% resources! Runs server callout scripts when *AN events occur! Q What are Ora#%e ata(ase (a#*$roun ro#esses se#i)i# to RAC
^F;SBloal Cache Service >rocess ^F;DBloal 7nueue Service Daemon ^F;ONBloal 7nueue Service ;onitor
^FCG'&nstance 7nueue >rocess "o ensure that each Oracle RAC dataase instance otains the loc- that it needs to satisf# a uer# or transaction. Oracle RAC instances use two processes. the Bloal Cache Service (BCS) and the Bloal 7nueue Service (B7S)! "he BCS and B7S maintain records of the statuses of each data file and each cached loc- using a Bloal Resource Director# (BRD)! "he BRD contents are distriuted across all of the active instances! Q What are Ora#%e C%usterware Co"onents
4oting isk Oracle RAC uses the voting dis- to manage cluster memership # wa# of a health chec- and aritrates cluster ownership among the instances in case of networ- failures! "he voting dis- must reside on shared dis-! Oracle luster /egistry %O/& ;aintains cluster configuration information as well as configuration information aout an# cluster dataase within the cluster! "he OCR must reside on shared dis- that is accessile # all of the nodes in #our cluster Q +ow o !ou trou(%eshoot noe re(oot
>lease chec- metalin- !!! Note 2[+[_!1 "rouleshooting CRS Reoots Note!++_[+!1 Hsing Diagwait as a diagnostic to get more information for diagnosing Oracle Clusterware Node evictions! Q +ow o !ou (a#*u the OCR
"here is an automatic ac-up mechanism for OCR! "he default location is ? :ORA,CRS,PO;7McdataM@clustername@M "o displa# ac-ups ? Jocrconfig =showac-up "o restore a ac-up ? Jocrconfig =restore /ith Oracle RAC 1'g Release 2 or later. #ou can also use the e%port command? Jocrconfig =e%port =s online. and use =import option to restore the contents ac-! /ith Oracle RAC 11g Release 1. #ou can do a manaual ac-up of the OCR with the command? J ocrconfig =manualac-up Q +ow o !ou (a#*u votin$ is*
Jdd ifLvoting,dis-,name ofLac-up,file,name Q +ow o I ienti)! the votin$ is* %o#ation
Jcrsctl uer# css votedis- Q +ow o I ienti)! the OCR )i%e %o#ation
chec-
@ssh@ are not reuired for normal Oracle RAC operation! Powever @ssh@ should e enaled for Oracle RAC and patchset installation! Q What is SCAN?
Single Client Access Name (SCAN) is s a new Oracle Real Application Clusters (RAC) 11g Release 2 feature that provides a single name for clients to access an Oracle Dataase running in a cluster! "he enefit is clients using SCAN do not need to change if #ou add or remove nodes in the cluster! Q What is the urose o) .rivate Inter#onne#t ?
Clusterware uses the private interconnect for cluster s#nchroni$ation (networ- hearteat) and daemon communication etween the the clustered nodes! "his communication is ased on the "C> protocol! RAC uses the interconnect for cache fusion (HD>) and inter=process communication ("C>)! Cache *usion is the remote memor# mapping of Oracle uffers. shared etween the caches of participating nodes in the cluster! Q Wh! o we have a Virtua% I. FVI.0 in Ora#%e RAC?
/ithout using &>s or *AN. clients connected to a node that died will often wait for a "C> timeout period (which can e up to 1' min) efore getting an error! As a result. #ou don0t reall# have a good PA solution without using &>s! /hen a node fails. the &> associated with it is automaticall# failed over to some other node and new node re=arps the world indicating a new ;AC address for the &>! Suseuent pac-ets sent to the &> go to the new node. which will send error RS" pac-ets ac- to the clients! "his results in the clients getting errors immediatel#
Q What o !ou o i) !ou see GC CR E9OCK 9OST in to Ti"e @vents in AWR Reort?
"his is most li-el# due to a fault in interconnect networ-! Chec- netstat =s if #ou see @fragments dropped@ or @pac-et reassemlies failed@ . /or- with #our s#stem administrator find the fault with networ-! Q +ow "an! noes are suorte in a RAC Data(ase?
1'g Release 2. support 1'' nodes in a cluster using Oracle Clusterware. and 1'' instances in a RAC dataase! Q Srv#t% #annot start instan#e3 I $et the )o%%owin$ error .RK.844 CRS8413 however s'%%us #an start it on (oth noes? +ow o !ou ienti)! the ro(%e"?
Set the environmental variale SR;,"RAC7 to true!! And start the instance with srvctl! Now #ou will get detailed error stac-! Q what is the urose o) the ONS ae"on?
"he Oracle Notification Service (ONS) daemon is an daemon started # the CRS clusterware as part of the nodeapps! "here is one ons daemon started per clustered node! "he Oracle Notification Service daemon receive a suset of pulished clusterware events via the local evmd and racgimon clusterware daemons and forward those events to application suscriers and to the local listeners! "his in order to facilitate? a! the *AN or *ast Application Notification feature or allowing applications to respond to dataase state changes! ! the 1'gR2 Foad 4alancing Advisor#. the feature that permit load alancing accross different rac nodes dependent of the load on the different nodes! "he rdms ;;ON is creating an advisor# for distriution of wor- ever# 'seconds and forward it via racgimon and ONS to listeners and applications! Q +ow o users #onne#t to ata(ase in an Ora#%e RAC environ"ent?
Hsers can access a RAC dataase using a client
Applications should use the services feature to connect to the Oracle dataase! Services enale
us to define rules and characteristics to control how users and applications connect to dataase instances! Q What are the #hara#teristi#s #ontro%%e (! Ora#%e servi#es )eature?
"he characteristics include a uniue name. wor-load alancing and failover options. and high availailit# characteristics! Q What is a votin$ is*?
A voting dis- is a file that manages information aout node memership! Q What are the a"inistrative tas*s invo%ve with votin$ is*?
*ollowing administrative tas-s are performed with the voting dis- ? 1) 4ac-ing up voting dis-s 2) Recovering oting dis-s ) Adding voting dis-s 3) Deleting voting dis-s +) ;oving voting dis-s Q +ow o we (a#*u votin$ is*s?
1) Oracle recommends that #ou ac- up #our voting dis- after the initial cluster creation and after we complete an# node addition or deletion procedures! 2) *irst. as root user. stop Oracle Clusterware (with the crsctl stop crs command) on all nodes! "hen. determine the current voting dis- # issuing the following command? crsctl uer# votedis- css ) "hen. issue the dd or ocop# command to ac- up a voting dis-. as appropriate! Bive the s#nta% of ac-ing up voting dis-s?= On Linux or UNIX systems? dd ifLvoting,dis-,name ofLac-up,file,name
where. voting_disk_name is the name of the active voting dis- backup_file_name is the name of the file to which we want to ac- up the voting dis- contents
On Windows systems. use the ocop# command? ocop# voting,dis-,name ac-up,file,name Q What is the Ora#%e Re#o""enation )or (a#*in$ u votin$ is*?
Oracle recommends us to use the dd command to ac-up the voting dis- with a minimum locsi$e of 3G4! Q +ow o !ou restore a votin$ is*?
"o restore the ac-up of #our voting dis-. issue the dd or ocop# command for Finu% and HN&E s#stems or ocop# for /indows s#stems respectivel#! On Linux or UNIX systems? dd ifLac-up,file,name ofLvoting,dis-,name On Windows systems. use the ocop# command? ocop# ac-up,file,name voting,dis-,name where. backup_file_name is the name of the voting dis- ac-up file voting_disk_name is the name of the active voting dis-
Q +ow #an we a an re"ove "u%ti%e votin$ is*s?
&f we have multiple voting dis-s. then we can remove the voting dis-s and add them ac- into our environment using the following commands. where path is the complete path of the location where the voting dis- resides? crsctl delete css votedisk path crsctl add css votedisk path
Q +ow o we sto Ora#%e C%usterware?When o we sto it?
4efore ma-ing an# modification to the voting dis-. as root user. stop Oracle Clusterware using the crsctl stop crs command on all nodes! Q +ow o we a votin$ is*?
"o add a voting dis-. issue the following command as the root user. replacing the path variale with the full# ualified path name for the voting dis- we want to add? crsctl add css votedis- path =force Q +ow o we "ove votin$ is*s?
"o move a voting dis-. issue the following commands as the root user. replacing the path variale with the full# ualified path name for the voting dis- we want to move? crsctl delete css votedis- path =force crsctl add css votedis- path =force Q +ow o we re"ove votin$ is*s?
"o remove a voting dis-. issue the following command as the root user. replacing the path variale with the full# ualified path name for the voting dis- we want to remove? crsctl delete css votedis- path =force Q What shou% we o a)ter "oi)!in$ votin$ is*s?
After modif#ing the voting dis-. restart Oracle Clusterware using the crsctl start crs command on all nodes. and verif# the voting dis- location using the following command? crsctl uer# css votedis- Q When #an we use 8)or#e otion?
&f our cluster is down. then we can include the =force option to modif# the voting disconfiguration. without interacting with active Oracle Clusterware daemons! Powever. using the =force option while an# cluster node is active ma# corrupt our configuration!
