| Oracle® Collaboration Suite High Availability Guide 10gRelease 1 (10.1.2) for Windows or UNIX Part Number B25481-03 |
|
|
View PDF |
| Oracle® Collaboration Suite High Availability Guide 10gRelease 1 (10.1.2) for Windows or UNIX Part Number B25481-03 |
|
|
View PDF |
This chapter describes how to manage the high availability environments in Oracle Collaboration Suite. Managing high availability environments primarily consists of managing nodes of Oracle Real Application Clusters and Oracle Collaboration Suite Applications. The chapter consists of the following sections:
Adding or Removing Oracle Collaboration Suite Applications Nodes
Configuring Manual Cold Failover for the Oracle Calendar Server
Starting and Stopping Oracle Collaboration Suite in a High Availability Environment
|
Note: For information about adding and removing nodes and instances in Oracle Real Application Clusters Database, refer to Chapter 5 in Real Application Clusters Administrator's Guide 10g Release 1 (10.1). |
This section describes the steps required in Oracle Collaboration Suite after adding or removing nodes from Oracle Collaboration Suite Database. The steps are as follows:
Step 2: Modify the Connect String of the Crawler Through the Search Admin Application
Step 3: Bounce the Oracle Collaboration Suite Application Tier Processes
The step to modify the Oracle Real Application Clusters database connect string in Oracle Internet Directory to include the new database instance node or remove the deleted database instance node is as follows:
In Oracle Internet Directory, navigate to dn:cn=OCSDB,cn=OracleContext and modify orclnetdesctring accordingly. This implies that if you add a database node, then add the new node specification to the connect string and if you delete a node, then remove the node specification from the connect string. An example for adding a database instance (this would all be on a single line in oidadmin but is split for legibility)is as follows.
If a third database instance had been added to a new node with the database name of ocsdb, dbnode3vip1.us.oracle.com, then you would modify dn:cn=OCSDB,cn=OracleContext orclnetdesctring value as follows ( the line to be added is shown in bold in the following example):
(DESCRIPTION=(ADDRESS_
LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=dbnode1vip1.us.oracle.com)(PORT=1521))
(ADDRESS=(PROTOCOL=TCP)
(HOST=dbnode2vip1.us.oracle.com)(PORT=1521))
(HOST=dbnode3vip1.us.oracle.com)(PORT=1521))
(LOAD_BALANCE=yes))(CONNECT_DATA=(SERVICE_NAME=ocsdb.us.oracle.com)))
When Oracle Real Application Clusters uses the Cluster File System (CFS), the Oracle Collaboration Suite Search crawler can be started from any of the Oracle Real Application Clusters nodes, as long as at least one Oracle Real Application Clusters node is up and running. When Oracle Real Application Clusters is not using CFS, the Oracle Collaboration Suite Search crawler always runs on a specified node. If this node stops operating, then you must run the wk0reconfig.sql script to move Oracle Collaboration Suite Search to another Oracle Real Application Clusters node. In the wk0reconfig.sql script, instance_name is the name of the Oracle Real Application Clusters instance that Oracle Collaboration Suite Search uses for crawling. After connecting to the database, to get the name of the current instance, run the following command:
SELECT instance_name FROM v$instance
connect_url is the JDBC connection string that guarantees a connection only to the specified instance. An example is as follows:
(DESCRIPTION=
(ADDRESS_LIST=
(ADDRESS=(PROTOCOL=TCP)
(HOST=<nodename>)
(PORT=<listener_port>)))
(CONNECT_DATA=(SERVICE_NAME=<service_name>)))
For versions before 10.1.2.2, run the following commands from the Oracle Collaboration Suite Application tier.
$ORACLE_HOME/opmn/bin/opmnctl stopall $ORACLE_HOME/ocas/bin/ocasctl -stopall $ORACLE_HOME/opmn/bin/opmnctl startall $ORACLE_HOME/ocas/bin/ocasctl –start –t ochecklet –p 8020 –n 1 $ORACLE_HOME/ocas/bin/ocasctl –start –t ocas –p 8010 –n 5
For 10.1.2.2, run the following commands from the Oracle Collaboration Suite Application tier.
$ORACLE_HOME/opmn/bin/opmnctl stopall $ORACLE_HOME/ocas/bin/ocasctl -stopall $ORACLE_HOME/opmn/bin/opmnctl startall $ORACLE_HOME/ocas/bin/ocasctl –startall
The default ports are 8010 and 8020. The valid range is 8010 to 8020.
This section describes how to add and delete Oracle Collboration Suite Applications nodes. The topics in this section are:
The steps for removing nodes from multiple Oracle Collaboration Suite Applications tiers that have a load balancer at the front end are as follows:
Shut down all Oracle Collaboration Suite Applications on the node of the Oracle Collaboration Suite Applications tier that must be removed, as follows:
opmnctl stopall
After shutting down Oracle Collaboration Suite Applications, the load balancer routes all requests to the active nodes assuming that the load balancer monitors have been set up correctly.
Remove the node from the load-balancer virtual server node pool.
The steps for adding nodes from multiple Oracle Collaboration Suite Applications are as follows:
Follow the steps listed in the "Installing Subsequent Instance of Oracle Collaboration Suite 10g Applications" section in Chapter 9 in the following guides:
Add the new node to the load balancer configuration
In an Oracle Collaboration Suite high availability environment, the Oracle Calendar server can be set up in an active-passive configuration, also known as cold failover cluster configuration.
In an Oracle Calendar server cold failover cluster configuration, if an active node fails and if you want an automated cold failover for Oracle Calendar Server, then you must use vendor clusterware. Manual cold failover is done when the vendor clusterware is not installed. The steps for manual cold failover configuration are as follows:
Ensure that the Oracle Calendar server virtual IP address is inaccessible by pinging it. If it is still accessible, then it may not be necessary to execute a cold failover. Validate that the system is truly down and also that it requires a cold failover before proceeding. If it is necessary and the Oracle Calendar server virtual IP is still accessible, then bring down the virtual IP address on the node where it is active.
Register the Oracle Calendar Server Oracle home with the central inventory. From the Oracle Calendar Server Oracle home, run the following command at the prompt:
$ORACLE_HOME/oui/bin/runInstaller -attachHome ORACLE_HOME=<Oracle_Home_Location> ORACLE_HOME_NAME=<Oracle_Home_Name>
After attaching the Oracle home, you can verify the success of the operation by verifying the contents of the log files present in the <central_inventory>/logs directory. You can also view the contents of the inventory.xml file under <central_inventory>/ContentsXML directory to verify that the Oracle home is registered.
Mount the Oracle Calendar server shared storage containing the ORACLE_HOME of Oracle Calendar Server on the destination node.
Bring up the virtual IP address on the destination node.
Start the Oracle Calendar server.
This section describes how to stop and start Oracle Collaboration Suite in a high availability environment. Set up the environment for the correct ORACLE_HOME when you log on to each tier before the stop and start processes. A database ORACLE_HOME will include the ORACLE_SID.
The steps for stopping Oracle Collaboration Suite in a high availability environment are as follows:
Log on to each Applications tier node and run the following commands:
$OH/opmn/bin/opmnctl stopall $OH/ocas/bin/ocasctl -stopall $OH/bin/emctl stop iasconsole $OH/bin/lsnrctl stop listener_es
For the Oracle Calendar server node, run the following commands:
$OH/opmn/bin/opmnctl stopall $OH/bin/emctl stop iasconsole
For each of the Identity Management tiers, do the following:
$OH/opmn/bin/opmnctl stopall $OH/bin/emctl stop iasconsole
In each Oracle Real Application Clusters database tier, run the following command:
emctl stop dbconsole
From any Oracle Real Application Clusters database node, run the command:
$OH/bin/srvctl stop database -d $dbname $OH/bin/svrctl stop nodeapps -n $nodename
In the preceding commands, $dbname is the name of the database and $nodename is the node name.
The steps for starting Oracle Collaboration Suite in a high availability environment are as follows:
In each Oracle Real Application Clusters database tier, run the following command:
emctl start dbconsole
From any Oracle Real Application Clusters database node, run the command:
$OH/bin/svrctl start nodeapps -n $nodename $OH/bin/srvctl start database -d $dbname
In the preceding commands, $dbname is the name of the database and $nodename is the node name.
Log on to each Identity Management tier to do the following:
$OH/opmn/bin/opmnctl startall $OH/bin/emctl start iasconsole
In a Distributed Identity Management Architecture, ensure that the Oracle Internet Directory tier is started before the OracleAS Single Sign-On tier.
The Oracle Calendar server is set up separately from the Applications tier in its own ORACLE_HOME. For the Oracle Calendar server node, ensure that the virtual IP is up and run the following commands:
$OH/opmn/bin/opmnctl startall $OH/bin/emctl start iasconsole
For each of the Applications tier instances, log in as root to start the listener_es process, ensure that port 25 is not in use and then run the following commands.
$OH/bin/tnslsnr listener_es -user $uid -group $gid
Log in as the owner of the Oracle Collaboration Suite software and perform the following steps for versions before 10.1.2.2:
$OH/ocas/bin/ocasctl -start -t ochecklet -p 8020 -n 1 $OH/ocas/bin/ocasctl -start -t ocas -p 8010 -n 5 $OH/opmn/bin/opmnctl startall $OH/bin/emctl start iasconsole
Log in as the owner of the Oracle Collaboration Suite software and perform the following steps for 10.1.2.2:
$OH/ocas/bin/ocasctl -startall $OH/opmn/bin/opmnctl startall $OH/bin/emctl start iasconsole
|
Note: This step assumes that you are using the default port (port 25) for the Oracle Mail listener. On UNIX systems, port 25 is a privileged port, so you must log in asroot. If you have installed the listener on a non-privileged port, then you do not need to log in as root. On Windows systems, you do not need to log in as root. |
