oel - Page 3 - Developer IT

Configure PERL DBI and DBD in Linux

- by Balualways

I am new to Perl and I work in a Linux OEL 5x server. I am trying to configure the Perl DB modules for Oracle connectivity (DBD and DBI modules). Can anyone help me out in the installation procedure? I had tried CPAN didn't really worked out. Any help would be appreciated. I am not quite sure I need to initialize any variables other than $LD_LIBRARY_PATH and $ORACLE_HOME These are my observations: ISSUE:: I am getting the following issue while using the DBI module to connect to Oracle: install_driver(Oracle) failed: Can't locate loadable object for module DBD::Oracle in @INC (@INC contains: /usr/lib64/perl5/site_perl/5.8.8/x86_64-linux-thread-multi /usr/lib/perl5/site_perl/5.8.8 /usr/lib/perl5/site_perl /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi /usr/lib/perl5/vendor_perl/5.8.8 /usr/lib/perl5/vendor_perl /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi /usr/lib/perl5/5.8.8 .) at (eval 3) line 3 Compilation failed in require at (eval 3) line 3. Perhaps a module that DBD::Oracle requires hasn't been fully installed at connectdb.pl line 57 I had installed the DBD for oracle from /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/DBD/DBD-Oracle-1.50 Could you please take a look into the steps and correct me if I am wrong: Observations: $ echo $LD_LIBRARY_PATH /opt/CA/UnicenterAutoSysJM/autosys/lib:/opt/CA/SharedComponents/Csam/SockAdapter/lib:/opt/CA/SharedComponents/ETPKI/lib:/opt/CA/CAlib $ echo $ORACLE_HOME /usr/local/oracle/ORA This is how I tried to install the DBD module: Download the file DBD 1.50 for Oracle Copy to /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/DBD Untar and Makefile.PL . Message: Using DBI 1.52 (for perl 5.008008 on x86_64-linux-thread-multi) installed in /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/DBI/ Configuring DBD::Oracle for perl 5.008008 on linux (x86_64-linux-thread-multi) Remember to actually *READ* the README file! Especially if you have any problems. Installing on a linux, Ver#2.6 Using Oracle in /opt/oracle/product/10.2 DEFINE _SQLPLUS_RELEASE = "1002000400" (CHAR) Oracle version 10.2.0.4 (10.2) Found /opt/oracle/product/10.2/rdbms/demo/demo_rdbms.mk Found /opt/oracle/product/10.2/rdbms/demo/demo_rdbms64.mk Found /opt/oracle/product/10.2/rdbms/lib/ins_rdbms.mk Using /opt/oracle/product/10.2/rdbms/demo/demo_rdbms.mk Your LD_LIBRARY_PATH env var is set to '/usr/local/oracle/ORA/lib:/usr/dt/lib:/usr/openwin/lib:/usr/local/oracle/ORA/ows/cartx/wodbc/1.0/util/lib:/usr/local/oracle/ORA/lib:/usr/local/sybase/OCS-12_0/lib:/usr/local/sybase/lib:/home/oracle/jdbc/jdbcoci73/lib:./' WARNING: Your LD_LIBRARY_PATH env var doesn't include '/opt/oracle/product/10.2/lib' but probably needs to. Reading /opt/oracle/product/10.2/rdbms/demo/demo_rdbms.mk Reading /usr/local/oracle/ORA/rdbms/lib/env_rdbms.mk Attempting to discover Oracle OCI build rules sh: make: command not found by executing: [make -f /opt/oracle/product/10.2/rdbms/demo/demo_rdbms.mk build ECHODO=echo ECHO=echo GENCLNTSH='echo genclntsh' CC=true OPTIMIZE= CCFLAGS= EXE=DBD_ORA_EXE OBJS=DBD_ORA_OBJ.o] WARNING: Oracle build rule discovery failed (32512) Add path to make command into your PATH environment variable. Oracle oci build prolog: [sh: make: command not found] Oracle oci build command: [] WARNING: Unable to interpret Oracle build commands from /opt/oracle/product/10.2/rdbms/demo/demo_rdbms.mk. (Will continue by using fallback approach.) Please report this to [email protected]. See README for what to include. Found header files in /opt/oracle/product/10.2/rdbms/public. client_version=10.2 DEFINE= -Wall -Wno-comment -DUTF8_SUPPORT -DORA_OCI_VERSION=\"10.2.0.4\" -DORA_OCI_102 Checking for functioning wait.ph System: perl5.008008 linux ca-build9.us.oracle.com 2.6.20-1.3002.fc6xen #1 smp thu apr 30 18:08:39 pdt 2009 x86_64 x86_64 x86_64 gnulinux Compiler: gcc -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic -D_REENTRANT -D_GNU_SOURCE -fno-strict-aliasing -pipe -Wdeclaration-after-statement -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -I/usr/include/gdbm Linker: not found Sysliblist: -ldl -lm -lpthread -lnsl -lirc Oracle makefiles would have used these definitions but we override them: CC: cc CFLAGS: $(GFLAG) $(OPTIMIZE) $(CDEBUG) $(CCFLAGS) $(PFLAGS)\ $(SHARED_CFLAG) $(USRFLAGS) [$(GFLAG) -O3 $(CDEBUG) -m32 $(TRIGRAPHS_CCFLAGS) -fPIC -I/usr/local/oracle/ORA/rdbms/demo -I/usr/local/oracle/ORA/rdbms/public -I/usr/local/oracle/ORA/plsql/public -I/usr/local/oracle/ORA/network/public -DLINUX -D_GNU_SOURCE -D_LARGEFILE64_SOURCE=1 -D_LARGEFILE_SOURCE=1 -DSLTS_ENABLE -DSLMXMX_ENABLE -D_REENTRANT -DNS_THREADS -fno-strict-aliasing $(LPFLAGS) $(USRFLAGS)] build: $(CC) $(ORALIBPATH) -o $(EXE) $(OBJS) $(OCISHAREDLIBS) [ cc -L$(LIBHOME) -L/usr/local/oracle/ORA/rdbms/lib/ -o $(EXE) $(OBJS) -lclntsh $(EXPDLIBS) $(EXOSLIBS) -ldl -lm -lpthread -lnsl -lirc -ldl -lm $(USRLIBS) -lpthread] LDFLAGS: $(LDFLAGS32) [-m32 -o $@ -L/usr/local/oracle/ORA/rdbms//lib32/ -L/usr/local/oracle/ORA/lib32/ -L/usr/local/oracle/ORA/lib32/stubs/] Linking with /usr/local/oracle/ORA/rdbms/lib/defopt.o -lclntsh -ldl -lm -lpthread -lnsl -lirc -ldl -lm -lpthread [from $(DEF_OPT) $(OCISHAREDLIBS)] Checking if your kit is complete... Looks good LD_RUN_PATH=/usr/local/oracle/ORA/lib Using DBD::Oracle 1.50. Using DBD::Oracle 1.50. Using DBI 1.52 (for perl 5.008008 on x86_64-linux-thread-multi) installed in /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/DBI/ Writing Makefile for DBD::Oracle Writing MYMETA.yml and MYMETA.json *** If you have problems... read all the log printed above, and the README and README.help.txt files. (Of course, you have read README by now anyway, haven't you?)

Read the article

Begin the Clone Wars Have!

- by Antony Reynolds

Creating a New Virtual Machine from an Existing Virtual Disk In previous posts I described how I set up an OEL6 machine under VirtualBox that can run an 11gR2 database and FMW 11.1.1.5. That is great if you want the DB and FMW running in the same virtual image and it has served me well for some proof of concepts and also for some testing of different JVMs. However I also wanted to run some testing of FMW with the database running on a separate physical machine. So in this post I will show how to take a VirtualBox image and create a new image based on the disks from that original image. What are my Options? There is more than one way to skin a cat, or in this case to create two separate VMs that can run on different hardware. Some of the options include: Create new virtual disk images for each new VM. Clone the existing disk images and point the new VM at the cloned images. Point the new VM at the existing snapshots. #1 is too much like hard work, install OEL twice, install a database again, install FMW again, run RCU again! Life is too short! #2 is probably the safest way of doing things. VirtualBox allows you to clone a disk image for use in a separate machine. However this of course duplicates the disk and means that it is now occupying 3 times the space, once for the original disk and twice more for the two clones I would need. #3 is the most space efficient way of doing things. It does mean however that I can only run the new “cloned” images if I have access to the original image because that is where the base snapshots reside. However this is not a problem for me as long as I remember to keep all threee images together. So this is the approach we will follow. Snapshot, What Snapshot? As we are going to create new virtual machines based on existing snapshots we need to figure out which snapshot to use. We do this by opening the “Media Manager” from within VirtualBox and moving the mouse over the snapshot images until we find the snapshots we want – the snapshot name is identified in the “Attached to:” comment. In my case I wanted the FMW installed snapshot because that had a database configured for FMW alongside the FMW software. I made a note of the filename of that snapshot (actually I just noted the first 5 characters as that was all that was needed to uniquely identify the snapshot file). When we create the new machines we will point them at the snapshot filename we have just checked. Network or NotWork? Because we want the two new machines to communicate with each other when hosted in different physical machines we can’t use the default NAT networking mode without a lot of hassle. But at the same time we need them to have fixed IP addresses relative to each other so that they can see each other whilst also being able to see the outside world. To achieve all these requirements I created two network adapters for each machine. Adapter 1 was a standard NAT mapping. This will allow each machine to get a dynamic IP address (10.0.2.15 by default) that can be used to access the external world through the VBox provided NAT gateway. This is the same as the existing configuration. The second adapter I created as a bridged adapter. This gives the virtual machine direct access to the host network card and by using fixed IP addresses each machine can see the other. It is important to choose fixed IP addresses that are not routable across your internal network so you don’t get any clashes with other machines on your network. Of course you could always get proper fixed IP addresses from your network people, but I have serveral people using my images and as long as I don’t have two instances of the same VM on the same network segment this is easier and avoids reconfiguring the network every time someone wants a copy of my VM. If it is available I would suggest using the 10.0.3.* network as 10.0.2.* is the default NAT network. You can check availability by pinging 10.0.3.1 and 10.0.3.2 from your host machine. If it times out then you are probably safe to use that. Creating the New VMs Now that I had collected the data that I needed I went ahead and created the new VMs. When asked for a “Boot Hard Disk” I used the “Choose a virtual hard disk file…” link to find the snapshot I had previously selected and set that to be the existing hard disk. I chose the previously existing SOA 11.1.1.5 install for both the new DB and FMW machines because that snapshot had the database with the RCU completed that I wanted for my DB machine and it had the SOA software installed which I wanted for my FMW machine. After the initial creation of the virtual machine go into the network setting section and enable a second adapter which will be bridged. Make a note of the MAC addresses (the last four digits should be sufficient) of the two adapters so that you can later set the bridged adapter to use fixed IP and the NAT adapter to use DHCP. We are now ready to start the VMs and reconfigure Linux. Reconfiguring Linux Because I now have two new machines I need to change their network configuration. In particular I need to change the hostname, update the hosts file and change the network settings. Changing the Hostname I renamed both hosts by running the hostname command as root: hostname vboxfmw.oracle.com I also edited the /etc/sysconfig file and set the correct hostname in there. HOSTNAME=vboxfmw.oracle.com Changing the Network Settings I needed to change the network configuration to give the bridged network a fixed IP address. I first explicitly set the MAC addresses of the two adapters, because the order of the virtual adapters in the VirtualBox Manager is not necessarily the same as the order of the adapters in the guest OS. So I went in to the System->Preferences->Network Connections screen and explicitly set the “Device MAC address” for the two adapters. Having correctly mapped the Linux adapters to the VirtualBox adapters I then set the Bridged adapter to use fixed IP addressing rather than DHCP. There is no need for additional routing or default gateways because we expect the two machine to be on the same LAN segment. Updating the Hosts File Having renamed the machines and reconfigured the network I then updated the /etc/hosts file to refer to the new machine name add a new line to the hosts file to provide an additional IP address for my server (the new fixed IP address) add a new line for the fixed IP address of the other virtual machine 10.0.3.101 vboxdb.oracle.com vboxdb # Added by NetworkManager 10.0.2.15 vboxdb.oracle.com vboxdb # Added by NetworkManager 10.0.3.102 vboxfmw.oracle.com vboxfmw # Added by NetworkManager 127.0.0.1 localhost.localdomain localhost ::1 vboxdb.oracle.com vboxdb localhost6.localdomain6 localhost6 To make sure everything takes effect I restarted the server. Reconfiguring the Database on the DB Machine Because we changed the hostname the listener and the EM console no longer start so I need to modify the listener.ora to use the new hostname and I also need to rebuild the EM configuration because it also relies on the hostname. I edited the $ORACLE_HOME/network/admin/listener.ora and changed the listening address to the new hostname: (ADDRESS = (PROTOCOL = TCP)(HOST = vboxdb.oracle.com)(PORT = 1521)) After changing the listener.ora I was able to start the listener using: lsnrctl start I also had to reconfigure the EM database control. I first deconfigured it using the command: emca -deconfig dbcontrol db -repos drop This drops the repository and removes any existing registered dbcontrols. I then re-configured it using the following command: emca -config dbcontrol db -repos create This creates the EM repository and then configures and starts dbcontrol. Now my database machine is ready so I can close it down and take a snapshot. Disabling the Database on the FMW Machine I set up the database to start automatically by creating a service called “dbora”. On the FMW machine I do not need the database running so I can prevent it auto-starting by running the following command: chkconfig –del dbora Note that because I am using a snapshot it is not a waste of disk space to have the DB installed but not used. As long as I don’t run it, it won’t cost me anything. I can now close the FMW machine down and take a snapshot. Creating a New Domain The FMW machine is now ready to create a new domain. When creating the domain I can point it at the second machine which is running the database. I can potentially run these machines on two separate physical machines as long as I have the original virtual machine available to both of the physical machines. Gotchas in Snapshotting VirtualBox does not support the concept of linked machines in a network like some virtualization technologies so when creating a snapshot it is a good idea to shut both VMs down and then take a snapshot on both of them. This is because we want to keep the database in sync with the middleware. One way to make sure that this happens would be to place all the domain configuration files on the database server via an NFS share, this would mean that all we would need to snapshot would be the database machine because that would hold all the state and configuration. The Sky’s the Limit We have covered a simple case of having just two machines. I have a more complicated configuration in which two machine run a RAC database off the same base OS image, and two more machines run a SOA cluster based on the same OS image. Just remember what machine holds state and what are the consequences of taking a snapshot.

Read the article

Clusterware 11gR2 – Setting up an Active/Passive failover configuration

- by Gilles Haro

Oracle is providing a large range of interesting solutions to ensure High Availability of the database. Dataguard, RAC or even both configurations (as recommended by Oracle for a Maximum Available Architecture - MAA) are the most frequently found and used solutions. But, when it comes to protecting a system with an Active/Passive architecture with failover capabilities, people often thinks to other expensive third party cluster systems. Oracle Clusterware technology, which comes along at no extra-cost with Oracle Database or Oracle Unbreakable Linux, is - in the knowing of most people - often linked to Oracle RAC and therefore, is seldom used to implement failover solutions. Oracle Clusterware 11gR2 (a part of Oracle 11gR2 Grid Infrastructure) provides a comprehensive framework to setup automatic failover configurations. It is actually possible to make "failover-able'", and then to protect, almost any kind of application (from the simple xclock to the most complex Application Server). Quoting Oracle: “Oracle Clusterware is a portable cluster software that allows clustering of single servers so that they cooperate as a single system. Oracle Clusterware also provides the required infrastructure for Oracle Real Application Clusters (RAC). In addition Oracle Clusterware enables the protection of any Oracle application or any other kind of application within a cluster.” In the next couple of lines, I will try to present the different steps to achieve this goal : Have a fully operational 11gR2 database protected by automatic failover capabilities. I assume you are fluent in installing Oracle Database 11gR2, Oracle Grid Infrastructure 11gR2 on a Linux system and that ASM is not a problem for you (as I am using it as a shared storage). If not, please have a look at Oracle Documentation. As often, I made my tests using an Oracle VirtualBox environment. The scripts are tested and functional on my system. Unfortunately, there can always be a typo or a mistake. This blog entry does not replace a course around the Clusterware Framework. I just hope it will let you see how powerful it is and that it will give you the whilst to go further with it... Note : This entry has been revised (rev.2) following comments from Philip Newlan. Prerequisite 2 Linux boxes (OELCluster01 and OELCluster02) at the same OS level. I used OEL 5 Update 5 with an Enterprise Kernel. Shared Storage (SAN). On my VirtualBox system, I used Openfiler to simulate the SAN Oracle 11gR2 Database (11.2.0.1) Oracle 11gR2 Grid Infrastructure (11.2.0.1) Step 1 - Install the software Using asmlib, create 3 ASM disks (ASM_CRS, ASM_DTA and ASM_FRA) Install Grid Infrastructure for a cluster (OELCluster01 and OELCluster02 are the 2 nodes of the cluster) Use ASM_CRS to store Voting Disk and OCR. Use SCAN. Install Oracle Database Standalone binaries on both nodes. Use asmca to check/mount the disk groups on 2 nodes Use dbca to create and configure a database on the primary node Let's name it DB11G. Copy the pfile, password file to the second node. Create adump directoty on the second node. Step 2 - Setup the resource to be protected After its creation with dbca, the database is automatically protected by the Oracle Restart technology available with Grid Infrastructure. Consequently, it restarts automatically (if possible) after a crash (ex: kill -9 smon). A database resource has been created for that in the Cluster Registry. We can observe this with the command : crsctl status resource that shows and ora.dba11g.db entry. Let's save the definition of this resource, for future use : mkdir -p /crs/11.2.0/HA_scripts chown oracle:oinstall /crs/11.2.0/HA_scripts crsctl status resource ora.db11g.db -p > /crs/11.2.0/HA_scripts/myResource.txt Although very interesting, Oracle Restart is not cluster aware and cannot restart the database on any other node of the cluster. So, let's remove it from the OCR definitions, we don't need it ! srvctl stop database -d DB11G srvctl remove database -d DB11G Instead of it, we need to create a new resource of a more general type : cluster_resource. Here are the steps to achieve this : Create an action script : /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh #!/bin/bash export ORACLE_HOME=/oracle/product/11.2.0/dbhome_1 export ORACLE_SID=DB11G case $1 in 'start') $ORACLE_HOME/bin/sqlplus /nolog <<EOF connect / as sysdba startup EOF RET=0 ;; 'stop') $ORACLE_HOME/bin/sqlplus /nolog <<EOF connect / as sysdba shutdown immediate EOF RET=0 ;; 'clean') $ORACLE_HOME/bin/sqlplus /nolog <<EOF connect / as sysdba shutdown abort ##for i in `ps -ef | grep -i $ORACLE_SID | awk '{print $2}' ` ;do kill -9 $i; done EOF RET=0 ;; 'check') ok=`ps -ef | grep smon | grep $ORACLE_SID | wc -l` if [ $ok = 0 ]; then RET=1 else RET=0 fi ;; '*') RET=0 ;; esac if [ $RET -eq 0 ]; then exit 0 else exit 1 fi This script must provide, at least, methods to start, stop, clean and check the database. It is self-explaining and contains nothing special. Just be aware that it must be runnable (+x), it runs as Oracle user (because of the ACL property - see later) and needs to know about the environment. Also make sure it exists on every node of the cluster. Moreover, as of 11.2, the clean method is mandatory. It must provide the “last gasp clean up”, for example, a shutdown abort or a kill –9 of all the remaining processes. chmod +x /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh scp /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh oracle@OELCluster02:/crs/11.2.0/HA_scripts Create a new resource file, based on the information we got from previous myResource.txt . Name it myNewResource.txt. myResource.txt is shown below. As we can see, it defines an ora.database.type resource, named ora.db11g.db. A lot of properties are related to this type of resource and do not need to be used for a cluster_resource. NAME=ora.db11g.db TYPE=ora.database.type ACL=owner:oracle:rwx,pgrp:oinstall:rwx,other::r-- ACTION_FAILURE_TEMPLATE= ACTION_SCRIPT= ACTIVE_PLACEMENT=1 AGENT_FILENAME=%CRS_HOME%/bin/oraagent%CRS_EXE_SUFFIX% AUTO_START=restore CARDINALITY=1 CHECK_INTERVAL=1 CHECK_TIMEOUT=600 CLUSTER_DATABASE=false DB_UNIQUE_NAME=DB11G DEFAULT_TEMPLATE=PROPERTY(RESOURCE_CLASS=database) PROPERTY(DB_UNIQUE_NAME= CONCAT(PARSE(%NAME%, ., 2), %USR_ORA_DOMAIN%, .)) ELEMENT(INSTANCE_NAME= %GEN_USR_ORA_INST_NAME%) DEGREE=1 DESCRIPTION=Oracle Database resource ENABLED=1 FAILOVER_DELAY=0 FAILURE_INTERVAL=60 FAILURE_THRESHOLD=1 GEN_AUDIT_FILE_DEST=/oracle/admin/DB11G/adump GEN_USR_ORA_INST_NAME= GEN_USR_ORA_INST_NAME@SERVERNAME(oelcluster01)=DB11G HOSTING_MEMBERS= INSTANCE_FAILOVER=0 LOAD=1 LOGGING_LEVEL=1 MANAGEMENT_POLICY=AUTOMATIC NLS_LANG= NOT_RESTARTING_TEMPLATE= OFFLINE_CHECK_INTERVAL=0 ORACLE_HOME=/oracle/product/11.2.0/dbhome_1 PLACEMENT=restricted PROFILE_CHANGE_TEMPLATE= RESTART_ATTEMPTS=2 ROLE=PRIMARY SCRIPT_TIMEOUT=60 SERVER_POOLS=ora.DB11G SPFILE=+DTA/DB11G/spfileDB11G.ora START_DEPENDENCIES=hard(ora.DTA.dg,ora.FRA.dg) weak(type:ora.listener.type,uniform:ora.ons,uniform:ora.eons) pullup(ora.DTA.dg,ora.FRA.dg) START_TIMEOUT=600 STATE_CHANGE_TEMPLATE= STOP_DEPENDENCIES=hard(intermediate:ora.asm,shutdown:ora.DTA.dg,shutdown:ora.FRA.dg) STOP_TIMEOUT=600 UPTIME_THRESHOLD=1h USR_ORA_DB_NAME=DB11G USR_ORA_DOMAIN=haroland USR_ORA_ENV= USR_ORA_FLAGS= USR_ORA_INST_NAME=DB11G USR_ORA_OPEN_MODE=open USR_ORA_OPI=false USR_ORA_STOP_MODE=immediate VERSION=11.2.0.1.0 I removed database type related entries from myResource.txt and modified some other to produce the following myNewResource.txt. Notice the NAME property that should not have the ora. prefix Notice the TYPE property that is not ora.database.type but cluster_resource. Notice the definition of ACTION_SCRIPT. Notice the HOSTING_MEMBERS that enumerates the members of the cluster (as returned by the olsnodes command). NAME=DB11G.db TYPE=cluster_resource DESCRIPTION=Oracle Database resource ACL=owner:oracle:rwx,pgrp:oinstall:rwx,other::r-- ACTION_SCRIPT=/crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh PLACEMENT=restricted ACTIVE_PLACEMENT=0 AUTO_START=restore CARDINALITY=1 CHECK_INTERVAL=10 DEGREE=1 ENABLED=1 HOSTING_MEMBERS=oelcluster01 oelcluster02 LOGGING_LEVEL=1 RESTART_ATTEMPTS=1 START_DEPENDENCIES=hard(ora.DTA.dg,ora.FRA.dg) weak(type:ora.listener.type,uniform:ora.ons,uniform:ora.eons) pullup(ora.DTA.dg,ora.FRA.dg) START_TIMEOUT=600 STOP_DEPENDENCIES=hard(intermediate:ora.asm,shutdown:ora.DTA.dg,shutdown:ora.FRA.dg) STOP_TIMEOUT=600 UPTIME_THRESHOLD=1h Register the resource. Take care of the resource type. It needs to be a cluster_resource and not a ora.database.type resource (Oracle recommendation) . crsctl add resource DB11G.db -type cluster_resource -file /crs/11.2.0/HA_scripts/myNewResource.txt Step 3 - Start the resource crsctl start resource DB11G.db This command launches the ACTION_SCRIPT with a start and a check parameter on the primary node of the cluster. Step 4 - Test this We will test the setup using 2 methods. crsctl relocate resource DB11G.db This command calls the ACTION_SCRIPT (on the two nodes) to stop the database on the active node and start it on the other node. Once done, we can revert back to the original node, but, this time we can use a more "MS$ like" method :Turn off the server on which the database is running. After short delay, you should observe that the database is relocated on node 1. Conclusion Once the software installed and the standalone database created (which is a rather common and usual task), the steps to reach the objective are quite easy : Create an executable action script on every node of the cluster. Create a resource file. Create/Register the resource with OCR using the resource file. Start the resource. This solution is a very interesting alternative to licensable third party solutions. References Clusterware 11gR2 documentation Oracle Clusterware Resource Reference Clusterware for Unbreakable Linux Using Oracle Clusterware to Protect A Single Instance Oracle Database 11gR1 (to have an idea of complexity) Oracle Clusterware on OTN Gilles Haro Technical Expert - Core Technology, Oracle Consulting

Read the article

Clusterware 11gR2 – Setting up an Active/Passive failover configuration

- by Gilles Haro

Oracle provides many interesting ways to ensure High Availability. Dataguard configurations, RAC configurations or even both (as recommended for a Maximum Available Architecture - MAA) are the most frequently found. But when it comes to protecting a system with an Active/Passive architecture with failover capabilities, one often thinks to expensive third party cluster systems. Oracle Clusterware technology, which comes free with Oracle Database, is – in the knowing of most people - often linked to Oracle RAC and therefore, is rarely used to implement failover solutions. 11gR2 Clusterware – which is part of Oracle Grid Infrastructure - provides a comprehensive framework to setup automatic failover configurations. It is actually possible to make “failover-able'” and, therefore to protect, almost every kind of application (from xclock to the more complex Application Server) In the next couple of lines, I will try to present the different steps to achieve this goal : Have a fully operational 11gR2 database protected by automatic failover capabilities. I assume you are fluent in installing Oracle Database 11gR2, Oracle Grid Infrastructure 11gR2 on a Linux system and that ASM is not a problem for you (as I am using it as a shared storage). If not, please have a look at Oracle Documentation. As often, I made my tests using an Oracle VirtualBox environment. The scripts are tested and functional. Unfortunately, there can always be a typo or a mistake. This blog entry is not a course around the Clusterware Framework. I just hope it will let you see how powerful it is and that it will give you the whilst to go further with it… Prerequisite 2 Linux boxes (OELCluster01 and OELCluster02) at the same OS level. I used OEL 5 Update 5 with Enterprise Kernel. Shared Storage (SAN). On my VirtualBox system, I used Openfiler to simulate the SAN Oracle 11gR2 Database (11.2.0.1) Oracle 11gR2 Grid Infrastructure (11.2.0.1) Step 1 – Install the software Using asmlib, create 3 ASM disks (ASM_CRS, ASM_DTA and ASM_FRA) Install Grid Infrastructure for a cluster (OELCluster01 and OELCluster02 are the 2 nodes of the cluster) Use ASM_CRS to store Voting Disk and OCR. Use SCAN. Install Oracle Database Standalone binaries on both nodes. Use asmca to check/mount the disk groups on 2 nodes Use dbca to create and configure a database on the primary node Let’s name it DB11G. Copy the pfile, password file to the second node. Create adump directoty on the second node. Step 2 - Setup the resource to be protected After its creation with dbca, the database is automatically protected by the Oracle Restart technology available with Grid Infrastructure. Consequently, it restarts automatically (if possible) after a crash (ex: kill –9 smon). A database resource has been created for that in the Cluster Registry. We can observe this with the command : crsctl status resource that shows and ora.dba11g.db entry. Let’s save the definition of this resource, for future use : mkdir –p /crs/11.2.0/HA_scripts chown oracle:oinstall /crs/11.2.0/HA_scripts crsctl status resource ora.db11g.db -p > /crs/11.2.0/HA_scripts/myResource.txt Although very interesting, Oracle Restart is not cluster aware and cannot restart the database on any other node of the cluster. So, let’s remove it from the OCR definitions, we don’t need it ! srvctl stop database -d DB11G srvctl remove database -d DB11G Instead of it, we need to create a new resource of a more general type : cluster_resource. Here are the steps to achieve this : Create an action script : /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh #!/bin/bash export ORACLE_HOME=/oracle/product/11.2.0/dbhome_1 export ORACLE_SID=DB11G case $1 in 'start') $ORACLE_HOME/bin/sqlplus /nolog <<EOF connect / as sysdba startup EOF RET=0 ;; 'stop') $ORACLE_HOME/bin/sqlplus /nolog <<EOF connect / as sysdba shutdown immediate EOF RET=0 ;; 'check') ok=`ps -ef | grep smon | grep $ORACLE_SID | wc -l` if [ $ok = 0 ]; then RET=1 else RET=0 fi ;; '*') RET=0 ;; esac if [ $RET -eq 0 ]; then exit 0 else exit 1 fi This script must provide, at least, methods to start, stop and check the database. It is self-explaining and contains nothing special. Just be aware that it is run as Oracle user (because of the ACL property – see later) and needs to know about the environment. It also needs to be present on every node of the cluster. chmod +x /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh scp /crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh oracle@OELCluster02:/crs/11.2.0/HA_scripts Create a new resource file, based on the information we got from previous myResource.txt . Name it myNewResource.txt. myResource.txt is shown below. As we can see, it defines an ora.database.type resource, named ora.db11g.db. A lot of properties are related to this type of resource and do not need to be used for a cluster_resource. NAME=ora.db11g.db TYPE=ora.database.type ACL=owner:oracle:rwx,pgrp:oinstall:rwx,other::r-- ACTION_FAILURE_TEMPLATE= ACTION_SCRIPT= ACTIVE_PLACEMENT=1 AGENT_FILENAME=%CRS_HOME%/bin/oraagent%CRS_EXE_SUFFIX% AUTO_START=restore CARDINALITY=1 CHECK_INTERVAL=1 CHECK_TIMEOUT=600 CLUSTER_DATABASE=false DB_UNIQUE_NAME=DB11G DEFAULT_TEMPLATE=PROPERTY(RESOURCE_CLASS=database) PROPERTY(DB_UNIQUE_NAME= CONCAT(PARSE(%NAME%, ., 2), %USR_ORA_DOMAIN%, .)) ELEMENT(INSTANCE_NAME= %GEN_USR_ORA_INST_NAME%) DEGREE=1 DESCRIPTION=Oracle Database resource ENABLED=1 FAILOVER_DELAY=0 FAILURE_INTERVAL=60 FAILURE_THRESHOLD=1 GEN_AUDIT_FILE_DEST=/oracle/admin/DB11G/adump GEN_USR_ORA_INST_NAME= GEN_USR_ORA_INST_NAME@SERVERNAME(oelcluster01)=DB11G HOSTING_MEMBERS= INSTANCE_FAILOVER=0 LOAD=1 LOGGING_LEVEL=1 MANAGEMENT_POLICY=AUTOMATIC NLS_LANG= NOT_RESTARTING_TEMPLATE= OFFLINE_CHECK_INTERVAL=0 ORACLE_HOME=/oracle/product/11.2.0/dbhome_1 PLACEMENT=restricted PROFILE_CHANGE_TEMPLATE= RESTART_ATTEMPTS=2 ROLE=PRIMARY SCRIPT_TIMEOUT=60 SERVER_POOLS=ora.DB11G SPFILE=+DTA/DB11G/spfileDB11G.ora START_DEPENDENCIES=hard(ora.DTA.dg,ora.FRA.dg) weak(type:ora.listener.type,uniform:ora.ons,uniform:ora.eons) pullup(ora.DTA.dg,ora.FRA.dg) START_TIMEOUT=600 STATE_CHANGE_TEMPLATE= STOP_DEPENDENCIES=hard(intermediate:ora.asm,shutdown:ora.DTA.dg,shutdown:ora.FRA.dg) STOP_TIMEOUT=600 UPTIME_THRESHOLD=1h USR_ORA_DB_NAME=DB11G USR_ORA_DOMAIN=haroland USR_ORA_ENV= USR_ORA_FLAGS= USR_ORA_INST_NAME=DB11G USR_ORA_OPEN_MODE=open USR_ORA_OPI=false USR_ORA_STOP_MODE=immediate VERSION=11.2.0.1.0 I removed database type related entries from myResource.txt and modified some other to produce the following myNewResource.txt. Notice the NAME property that should not have the ora. prefix Notice the TYPE property that is not ora.database.type but cluster_resource. Notice the definition of ACTION_SCRIPT. Notice the HOSTING_MEMBERS that enumerates the members of the cluster (as returned by the olsnodes command). NAME=DB11G.db TYPE=cluster_resource DESCRIPTION=Oracle Database resource ACL=owner:oracle:rwx,pgrp:oinstall:rwx,other::r-- ACTION_SCRIPT=/crs/11.2.0/HA_scripts/my_ActivePassive_Cluster.sh PLACEMENT=restricted ACTIVE_PLACEMENT=0 AUTO_START=restore CARDINALITY=1 CHECK_INTERVAL=10 DEGREE=1 ENABLED=1 HOSTING_MEMBERS=oelcluster01 oelcluster02 LOGGING_LEVEL=1 RESTART_ATTEMPTS=1 START_DEPENDENCIES=hard(ora.DTA.dg,ora.FRA.dg) weak(type:ora.listener.type,uniform:ora.ons,uniform:ora.eons) pullup(ora.DTA.dg,ora.FRA.dg) START_TIMEOUT=600 STOP_DEPENDENCIES=hard(intermediate:ora.asm,shutdown:ora.DTA.dg,shutdown:ora.FRA.dg) STOP_TIMEOUT=600 UPTIME_THRESHOLD=1h Register the resource. Take care of the resource type. It needs to be a cluster_resource and not a ora.database.type resource (Oracle recommendation) . crsctl add resource DB11G.db -type cluster_resource -file /crs/11.2.0/HA_scripts/myNewResource.txt Step 3 - Start the resource crsctl start resource DB11G.db This command launches the ACTION_SCRIPT with a start and a check parameter on the primary node of the cluster. Step 4 - Test this We will test the setup using 2 methods. crsctl relocate resource DB11G.db This command calls the ACTION_SCRIPT (on the two nodes) to stop the database on the active node and start it on the other node. Once done, we can revert back to the original node, but, this time we can use a more “MS$ like” method :Turn off the server on which the database is running. After short delay, you should observe that the database is relocated on node 1. Conclusion Once the software installed and the standalone database created (which is a rather common and usual task), the steps to reach the objective are quite easy : Create an executable action script on every node of the cluster. Create a resource file. Create/Register the resource with OCR using the resource file. Start the resource. This solution is a very interesting alternative to licensable third party solutions. References Clusterware 11gR2 documentation Oracle Clusterware Resource Reference Gilles Haro Technical Expert - Core Technology, Oracle Consulting

Read the article

Javascript: Can't control parent of descendant nodes.

- by .phjasper

I'm creating elements (level 1) dynamically which in turn create elements (level 2) themselves. However, the children of level 2 elements have "body" as their parent. In the HTML code below, the content if spotAd2 is created by my function createNode(). It's a Google Ad Sense tag. However, the Google Ad Sense tag create elements that went directly under "body". I need them to by under spotAd2. function createNode( t, // type. tn, // if type is element, tag name. a, // if type is element, attributes. v, // node value or text content p, // parent f ) // whether to make dist the first child or not. { n = null; switch( t ) { case "element": n = document.createElement( tn ); if( a ) { for( k in a ) { n.setAttribute( k, a[ k ] ); } } break; case "text": case "cdata_section": case "comment": n = document.createTextNode(v); break; } if ( p ) { if( f ) { p.insertBefore( n, p.firstChild ); } else { p.appendChild( n ); } } return n; } spotAd2 = document.getElementById("spotAd2"); n1 = createNode("element", "div", {"id":"tnDiv1"}, "\n" , null, true); n2 = createNode("element", "script", {"type":"text\/javascript"}, "\n" , n1, false); n3 = createNode("comment", "", null, "\n" + "google_ad_client = \"pub-0321943928525350\";\n" + "/* 728x90 (main top) */\n" + "google_ad_slot = \"2783893649\";\n" + "google_ad_width = 728;\n" + "google_ad_height = 90;\n" + "//\n" , n2, false); n4 = createNode("element", "script", {"type":"text\/javascript","src":"http:\/\/pagead2.googlesyndication.com\/pagead\/show_ads.js"}, "\n" , n1, false); --- Result: <body> <table cellspacing="2" cellpadding="2" border="1"> <tbody><tr> <td>Oel ngati kemeie</td> <td>Kamakto niwin</td> </tr> <tr> <td>The ad:</td> <td> <div id="spotAd2">  <div id="tnDiv1"> <script type="text/javascript"> google_ad_client = "pub-0321943928525350"; /* 728x90 (main top) */ google_ad_slot = "2783893649"; google_ad_width = 728; google_ad_height = 90; </script> <script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"></script> </div>  </div> </td> </tr> <tr> <td>txopu ra'a tsi, tsamsiyu</td> <td>teyrakup skxawng</td> </tr> </tbody></table>  <script src="http://pagead2.googlesyndication.com/pagead/expansion_embed.js"></script> <script src="http://googleads.g.doubleclick.net/pagead/test_domain.js"></script> <script>google_protectAndRun("ads_core.google_render_ad", google_handleError, google_render_ad);</script> <ins style="border: medium none ; margin: 0pt; padding: 0pt; display: inline-table; height: 90px; position: relative; visibility: visible; width: 728px;"> <ins style="border: medium none ; margin: 0pt; padding: 0pt; display: block; height: 90px; position: relative; visibility: visible; width: 728px;"> <iframe width="728" scrolling="no" height="90" frameborder="0" vspace="0" style="left: 0pt; position: absolute; top: 0pt;" src="http://googleads.g.doubleclick.net/pagead/ads?client=ca-pub-0321943928525350&output=html&h=90&slotname=2783893649&w=728&lmt=1273708979&flash=10.0.45&url=http%3A%2F%2Fkenshin.katanatechworks.com%2Ftest%2FadsBrowserSide.php&dt=1273708980294&shv=r20100422&correlator=1273708980298&frm=0&ga_vid=695691836.1273708981&ga_sid=1273708981&ga_hid=1961182006&ga_fc=0&u_tz=480&u_his=2&u_java=1&u_h=1080&u_w=1920&u_ah=1052&u_aw=1920&u_cd=24&u_nplug=5&u_nmime=38&biw=1394&bih=324&fu=0&ifi=1&dtd=955&xpc=Jl67G4xiq6&p=http%3A//kenshin.katanatechworks.com" name="google_ads_frame" marginwidth="0" marginheight="0" id="google_ads_frame1" hspace="0" allowtransparency="true"> </iframe> </ins> </ins>  </body>

Developer IT