|
Contents
|
Components of the N1 SPS Oracle 10g R2 RAC Plug-In
The N1 SPS Oracle 10g R2 RAC plug-in is made up of a large number of components and component types. These components are described in the following sections together with the rules for their use.
Plans in the N1 SPS Oracle 10g R2 RAC Plug-In covers the targeting requirements for these components.
There are additional base component types that are created by the plug-in, but these are not exposed to the user through any of the plans. No attempt should be made to use these directly.
Standard Component Installation/Expansion and Contraction/Removal Semantics
The Oracle 10g R2 RAC plug-in is designed to accommodate installations, removals, and deployment changes that are made outside the N1 SPS environment. For example, you might have installed Oracle Clusterware and RAC on a system, but now you want to expand that deployment to an additional cluster node. By using the capabilities of the plug-in, you can first attach to the existing deployment and then grow or shrink it through the plug-in.
Each component has a single installation and a single uninstallation block. The installation steps check to see whether the requested configuration for the component is already in place, such as, has this Solaris package already been installed, or has this Solaris user already been created. If the object already exists and the configuration matches (within the limits of the component checks), then the installation performs no further action and completes successfully. If the object exists and can be expanded or modified to match the request then the operation is attempted. If the object exists but the requested configuration would require the installation to perform a contraction then the installation fails. Finally, if the object does not exist and the install completes successfully, then the object will be installed. This approach avoids having to have separate plans for installing an object and another to expanding it.
 | Note
See the individual component sections for details of the configuration checks that are performed. |
Component removal is designed to work in a similar way. For example, if a four-node cluster has Oracle Clusterware installed on all four nodes and you want to remove one of these nodes from the configuration, then you would target the node that you wanted to delete by using the removal plan. By targeting just one node, you are indicating that you only want to shrink the configuration, rather than remove it altogether. Alternatively, by targeting all four nodes with the same plan, the entire Clusterware deployment will be removed. This approach avoids the complexity of having separate plans for shrinking an object and for removing it.
Overall, the approach can be summarize by saying that when you install or expanding a component, you target all the nodes that you want to be in the final, or "To Be", state. And when shrinking or removing a component, you target only those nodes that you want to remove the object from.
This section describes the following semantics for each of the Sun Cluster resource group and resource component installation and removal plans:
Top
Sun Cluster Resource Group Install/Attach/Expand Semantics
When one of the Sun Cluster resource group component "create" plans is run, it tests to see whether the resource group exists already. If it does not exist, the resource group is created. If the resource group exists and its node list matches the nodeList property set by the component variable, then no further action is taken. This constitutes an attach. If it does exist, but the resource group node list property is a subset of the nodeList component variable, then the resource group is extended onto the new node. Finally, if the resource group node list property is a superset, or a disjoint set, of the nodeList component variable, then the component installation fails on all the nodes targeted.
Top
Sun Cluster Resource Group Contract/Remove Semantics
When one of the Sun Cluster resource group component "remove" plans is run, it checks to see what the resource group's node list property is set to. If the target node list matches or is a superset of the resource group node list, then the resource group is completely removed from the cluster, together with all the resources it contains. However, if the target nodes form a subset of the current resource group node list, then those nodes are removed from the resource group's node list.
Top
Sun Cluster Resource Install/Attach/Modify Semantics
When one of the Sun Cluster resource component "create" plans is run, it tests to see whether the node list of the resource group specified matches the nodes targeted. If it does not match, the plan fails. If it does match and the resource does not exist in that resource group, then the resource is created. If the resource does exist in the resource group, then the installation plan attempts to set the properties of the resource to those specified in the component variables. This final operation is considered an attach or a modify, depending on whether any properties are changed.
Top
Sun Cluster Resource Remove Semantics
When one of the Sun Cluster resource component "remove" plans is run, it tests to see whether the resource exists in the specified resource group. If it does not exist, the removal completes successfully. If it does exist and the target node list matches that of the resource group that contains it, then the resource is removed. If it exists but the target node list does not match the resource groups node list, then the operation fails. Finally, if the resource exists in the resource group with a node list: nodeA, nodeB and the plan has targeted nodeC and an N1 SPS resource object is present, then the N1 SPS resource component will be deleted. This is important when you want to bring the N1 SPS view of resource deployment into line with the actual Sun Cluster resource deployment after a resource group node list is shrunk. You would achieve this by first shrinking the resource group, and then by removing the resource component from the node which no longer has the resource group in its node list.
Example: Suppose you have an existing four-node (pgyruss1, pgyruss2 and pgyruss3, pgyruss4) Sun Cluster where you have used the plug-in to deploy all the Sun Cluster packages and components that are needed to support Oracle Clusterware. However, you have manually installed Oracle Clusterware on three nodes pgyruss1, pgyruss2 and pgyruss3 but you want to expand this to four nodes, adding pgyruss4 , and then later on shrink the deployment to just pgyruss1 and pgyruss4. How would you simplify this exercise by using the plug-in?
First, you need to attach to the existing Oracle Clusterware configuration. To do this, create a CRS component variable set that matches your current 3 node deployment. (Only the installPath and nodeList component variables are important because are used to check for existing deployments). You then deploy the component, targeting the three physical nodes that Clusterware is already on. The installation logic recognizes that the settings that you provided match the existing deployment and the component installs successfully, taking no further action.
Now create a new CRS component variable set with a longer nodeList that includes pgyruss4. When you deploy the component again, this time targeting all four nodes, the installation logic recognizes that the existing deployment is a subset of the nodes you have requested and therefore needs expanding. It then performs the steps required to achieve this state.
The preceding process can actually be achieved by using just the second task, but it serves to demonstrate how the CRS component can be deployed repeatedly to reach a desired configuration.
The final step of shrinking the deployment from four nodes to two uses the CRS removal plan, targeting nodes pgyruss2 and pgyruss3. When complete, it leaves just pgyruss1 and pgyruss4 with CRS installed. This is reflected in the N1 SPS component view of the configuration, too.
Synchronized and Nonsynchronized Component Installations
This section describes synchronized and nonsynchronized component installations.
Top
When a Sun Cluster resource group or resource is created, it is automatically visible, by using the /usr/cluster/bin/scstat -g or /usr/cluster/bin/clrg status command, from all cluster nodes. These objects also have a node list property which define the nodes on which they can be present on. When deploying these and other objects, such as Oracle Clusterware, RAC and the databases, the RAC plug-in is designed to reflect this state. This is achieved by targeting component installation at all of the hosts of the "To Be" state, as described previously. As some of the underlying creation or installation is singular in nature, such as a cluster resource group is created only on one node and yet visible from all, the installation steps have to be carefully synchronized to ensure that only the appropriate nodes perform the necessary steps. Consequently, some components have what are termed synchronized component installation while others have nonsynchronized installation, that is, ones that can be deployed individually, in any order and at any time. Broadly speaking, all Sun Cluster, Oracle Clusterware or RAC object, and shared QFS file system creations are synchronized, while all user creation and package installation is not. This section highlights which plans are synchronized and which are not.
Synchronized plans call the following two sub-plans before calling the component itself:
- Synchronization_Service.clear_up
- Synchronization_Service.register
When you run these types of plan, you must ensure that the "Use these hosts and/or host sets for all subplans" check box remains checked.
The synchronization mechanism uses the Sun Cluster global-devices file systems.
| Note
Do not attempt to run more than one synchronized plan against any cluster at any one time. Doing so might cause unpredictable results. |
You can view the installation, deinstallation, and control steps for the individual components in the N1 SPS browser by navigating to the /com/sun/oracle10g_RAC directory.
| Note
You must run all plans from the Oracle 10g RAC plug-in page. This ensures that the correct synchronization steps are performed. Failure to do so might cause unpredictable results. |
Solaris User Component
This section describes the Solaris user component, Oracle_User.
Top
Oracle User Component (Oracle_User)
The Oracle user is the most important component in the plug-in. The Oracle UDLM, CRS, RAC, and database components derive their own variables from the Oracle_User component variable settings. Therefore, it must be deployed to the target cluster nodes first, even if the user account already exists on a node. If the user does not exist, the password is set to newroot. The installation also adds all the existing cluster nodes to the .rhosts file in the Oracle user's home directory.
The Oracle user component defines the following component variables.
| Name |
Description |
Notes |
| installName |
The Oracle user name |
Mandatory |
| installPath |
The Oracle home directory |
Mandatory |
| softUserId |
The Oracle user id |
Mandatory |
| softGroup |
The Oracle user's primary group name |
Mandatory – default oinstall |
| softGroupId |
The Oracle user's primary group ID |
Mandatory |
| dbaGroup |
The Oracle user's secondary (dba) group name |
Mandatory – default dba |
| dbaGroupId |
The Oracle user's secondary (dba) group ID |
Mandatory |
| operGroup |
The Oracle user's secondary (operator) group name |
Mandatory but can be the same as dba group |
| operGroupId |
The Oracle user's secondary (operator) group ID |
Mandatory but can be the same as dba group |
| shell |
The shell to use |
Mandatory – must be a supported Solaris shell, such as /bin/sh, /bin/ksh, or /bin/csh |
If you do not need to configure an operator group, simply make the operGroup and operGroupId values the same as the dbaGroup and dbaGroupId component variable values, respectively.
Solaris Package Components
This section describes the following Solaris package components:
Top
Sun Cluster Reconfiguration Framework Package Component (SUNWscucm_package)
The Sun Cluster reconfiguration framework package (SUNWscucm) must be installed before you can create the required Sun Cluster resource groups and resources that are needed to support Oracle Clusterware. The package is not installed by default under Sun Cluster 3.1 or 3.2 software.
The component has the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath variable should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
For Sun Cluster 3.2 software, the installPath should be set to the platform specific directory, such as Solaris_sparc, that contains the installer executable on the Java Availability Suite distribution media. This directory also contains the Product subdirectory in which all the Sun Cluster packages are located.
For further information see Chapter 1, Installing Sun Cluster Support for Oracle RAC, in Sun Cluster Data Service for Oracle RAC Guide for Solaris OS.
Top
Sun Cluster UNIX Distributed Lock Manager Package Components (SUNWudlmr_package and SUNWudlm_package)
The Sun Cluster UNIX Distributed Lock Manager package (SUNWudlmr and SUNWudlm) must be installed on all SPARC based platforms before you can create the required Sun Cluster resource groups and resources that are needed to support Oracle Clusterware. The package is not installed by default under Sun Cluster 3.1 or 3.2 software.
The components each have the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath variable should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
For Sun Cluster 3.2 software, the installPath variable should be set to the platform specific directory, such as Solaris_sparc, that contains the installer executable on the Java Availability Suite distribution media. This directory also contains the Product subdirectory in which all the Sun Cluster packages are located.
Top
Oracle UNIX Distributed Lock Manager Package Component (ORCLudlm_package)
The Oracle UNIX Distributed Lock Manager package (ORCLudlm) must be installed on all SPARC based platforms before the required Sun Cluster resource groups and resources that are needed to support Oracle Clusterware can be created.
The components each have the following component variable.
| Name |
Description |
Notes |
| installPath |
Enter the directory that contains the Oracle UDLM package |
Mandatory |
A compressed tar file of the package (ORCLudlm.tar.Z) is stored in the racpatch subdirectory of the Oracle Clusterware release media. This must be uncompressed and extracted into a suitable location, such as /tmp.
Top
Sun Cluster Solaris Volume Manager Integration Package Component (SUNWscmd_package)
The Sun Cluster Solaris Volume Manager integration package (SUNWscmd) must be installed if you plan to use Solaris Volume Manager shared disksets for either shared QFS file systems or raw Oracle data. The package is not installed by default under Sun Cluster 3.1 or 3.2 software.
The component has the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
For Sun Cluster 3.2 software, the installPath should be set to the platform specific directory, such as Solaris_sparc, that contains the installer executable on the Java Availability Suite distribution media. This directory also contains the Product subdirectory in which all the Sun Cluster packages are located.
Top
Sun Cluster Veritas Cluster Volume Manager Integration Package Components (SUNWcvmr_package and SUNWcvm_package)
The Sun Cluster Veritas Volume Manager integration packages (SUNWcvmr and SUNWcvm) must be installed if you plan to use Veritas Volume Manager shared disk groups for raw Oracle data, such as the Oracle Configuration Repository (OCR), voting disks, redo logs, or tablespaces. The package is not installed by default under Sun Cluster 3.1 or 3.2 software.
The components each have the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
For Sun Cluster 3.2 software, the installPath should be set to the platform-specific directory, such as Solaris_sparc, that contains the installer executable on the Java Availability Suite distribution media. This directory also contains the Product subdirectory in which all the Sun Cluster packages are located.
Top
Sun Cluster Hardware RAID Integration Package Component (SUNWschwr_package)
The Sun Cluster hardware RAID integration package (SUNWschwr) must be installed if you plan to use hardware RAID LUNs without an additional volume manager for raw Oracle data, such as the Oracle Configuration Repository (OCR), voting disks, redo logs, or tablespaces. The package is not installed by default under Sun Cluster 3.1 software and does not exist under Sun Cluster 3.2 software.
The component has the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
Top
Sun Cluster Oracle Agent Package Component (SUNWscor_package)
The Sun Cluster Oracle agent package (SUNWscor) must be installed if you want to use the Oracle RAC integration features of Sun Cluster 3.2 software. The package is not installed by default under Sun Cluster 3.2 software.
The component has the following component variable.
| Name |
Description |
Notes |
| installPath |
The full path to the Java Availability Suite or Oracle RAC agents media installer directory, where the installer executable is located. |
Mandatory |
For Sun Cluster 3.1 software, the installPath should be set to the directory that contains the installer executable on the Sun Cluster 3.1 agents distribution media. This directory also contains the components subdirectory in which all the Sun Cluster agent packages are located.
For Sun Cluster 3.2 software, the installPath should be set to the platform specific directory, such as Solaris_sparc, that contains the installer executable on the Java Availability Suite distribution media. This directory also contains the Product subdirectory in which all the Sun Cluster packages are located.
Top
Sun StorageTek QFS Root and User Package Components (SUNWqfsr_package and SUNWqfsu_package)
The Sun StorageTek QFS packages must be installed if you want to create shared QFS file systems on the cluster.
The components each have the following component variable.
| Name |
Description |
Notes |
| installPath |
Provide the path to the SUNWqfsr or SUNWqfsu package |
Mandatory |
The installPath variable should be set to the directory that contains the SUNWqfsr or SUNWqfsu packages on the QFS distribution media. These are normally located under the architecture/solaris_release directory of the media.
File System Component
This section describes the file system component, Shared_QFS.
Top
Shared QFS File System Component (Shared_QFS)
The shared QFS file system component is used to create, extend new shared QFS file systems or attach to existing ones. The plug-in supports QFS version 4.5. Only ma file-system types are supported to use mm metadata devices and mr data devices.
The component has the following component variables.
| Name |
Description |
Notes |
| installName |
The QFS name for the file system |
Mandatory |
| installPath |
The mount point for the shared QFS file system |
Mandatory |
| mountOptions |
Provide a comma-separated list of mount options for the shared QFS file system. Shared is assumed |
Optional, shared is assumed. Default value is: forcedirectio,nosam,stripe=1, sync_meta=1,mh_write,qwrite, notrace,nstreams=1024, rdlease=600,wrlease=600, aplease=600 |
| baseId |
The base ID for entries to create in the mcf file. |
Mandatory |
| metadataDeviceList |
A comma-separated list of devices to use for the file system Meta information |
Mandatory |
| dataDeviceList |
A comma-separated list of devices to use for file system data |
Mandatory |
| DAU |
DAU (Allocation unit) in multiples of 1 KB |
Optional default is 16 KB |
See the QFS documentation for more details on the architecture of QFS.
Sun Cluster Resource Group Components
This section describes the following resource group components:
Top
RAC Framework Resource Group Component (RAC_framework_RG)
The RAC_framework_RG component enables you to create the Sun Cluster resource group that is suitable for holding the resources such as rac_framework, rac_svm, or rac_udlm that are needed to support an Oracle Clusterware or RAC installation. You will need to deploy this resource group before you create these other resources. A Sun Cluster configuration only has one RAC framework resource group, however, the component will not prevent you from attempting to deploy more than one.
For an overview of the work that the plug-in does, see Configuring Resources for Oracle RAC Database Instances in Sun Cluster Data Service for Oracle RAC Guide for Solaris OS.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource group name |
Mandatory – default is rac-framework-rg |
| maximum_primaries |
The maximum number of primaries for the resource group |
Mandatory – but if left blank, will default to the number of nodes in the node list |
| desired_primaries |
The desired number of primaries for the resource group |
Mandatory – but if left blank, will default to the number of nodes in the node list |
| nodeList |
The comma-separated node list for the resource group such as node1,node2,node3 |
Mandatory |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
RAC Proxy Resource Group Component (RAC_proxy_RG)
The RAC_proxy_RG component enables you to create a Sun Cluster resource group that is suitable for holding a RAC proxy resource, which is used to manage an Oracle RAC database. This feature is only available on installations that run Sun Cluster 3.2 software. You will need to deploy this resource group before you create RAC proxy resources. A Sun Cluster 3.2 installation can have multiple RAC proxy resource groups, so you can deploy this component multiple times to the same cluster.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource group name |
Mandatory – default is rac-server-proxy-rg |
| maximum_primaries |
The maximum number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| desired_primaries |
The desired number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| nodeList |
The comma-separated node list for the resource group such as node1,node2,node3 |
Mandatory |
| rg_affinities |
The resource group affinities in a comma-separated list |
Mandatory – should be set to rac-framework-rg,++name_of_scalable_dg_rg |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Scalable Device Group Resource Group Component (Scalable_Device_Group_RG)
This component enables you to create a Sun Cluster resource group that is suitable for holding a scalable device group resource, which is used check on the availability of the specified device groups, that is, Solaris Volume Manager metasets or Veritas Volume Manager disk groups.
This feature is only available on installations that run Sun Cluster 3.2 software. You will need to deploy this resource group before you create scalable mount points. A Sun Cluster 3.2 installation can have multiple scalable device group resource groups, so you can deploy this component multiple times to the same cluster.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource group name |
Mandatory |
| maximum_primaries |
The maximum number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| desired_primaries |
The desired number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| nodeList |
The comma-separated node list for the resource group such as node1,node2,node3 |
Mandatory |
| rg_affinities |
The resource group affinities |
Optional – defaults to ++rac-framework-rg |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Scalable Mount Point Resource Group Component (Scalable_Mount_Point_RG)
This component enables you to create a Sun Cluster resource group that is suitable for holding a scalable mount point resource, which is used to check on the availability of the specified mounted file systems, such as Shared QFS or Network Attached Storage (NAS). This feature is only available on installations that run Sun Cluster 3.2 software. Deploy this resource group before you create scalable device group resources. A Sun Cluster 3.2 installation can have multiple scalable mount point resource groups, so you can deploy this component multiple times to the same cluster.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource group name |
Mandatory |
| maximum_primaries |
The maximum number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| desired_primaries |
The desired number of primaries for the resource group |
Mandatory – but if left blank will default to the number of nodes in the node list |
| nodeList |
The comma-separated node list for the resource group such as node1,node2,node3 |
Mandatory |
| rg_affinities |
The resource group affinities |
Optional – should be set to ++name_of_scalable_dg-rg for shared QFS mount points |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Shared QFS Resource Group Component (Shared_QFS_RG)
This component enables you to create a Sun Cluster resource group that is suitable for holding a QFS metadata server resource. These resource groups are used to make the QFS metadata server highly available. You will need to deploy this resource group before you create a shared QFS resource. A Sun Cluster 3.1 or 3.2 installation can have multiple shared QFS resource groups, so you can deploy this component multiple times to the same cluster.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource group name |
Mandatory |
| nodeList |
The comma-separated node list for the resource group, such as node1,node2,node3 |
Mandatory |
| rg_affinities |
The resource group affinities |
Optional |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Sun Cluster Resource Components
This section describes the following resource components:
Top
RAC Framework Resource Component (RAC_framework_RS)
The RAC_framework_RS component enables you to create the Sun Cluster RAC framework resource that is needed to drive the cluster reconfiguration process that is used by Oracle Clusterware. Deploy this resource before you create the RAC_udlm (SPARC only) or volume manager, RAC_svm or RAC_cvm, resources. A Sun Cluster configuration has only one RAC framework resource and the component will prevent you from attempting to deploy more than one. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac-framework |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
Top
CRS Framework Resource Component (CRS_framework_RS)
The CRS_framework_RS component enables you to create the Sun Cluster CRS framework resource that is used to monitor and shutdown Oracle Clusterware. This feature is only available on installations that run Sun Cluster 3.2 software. A Sun Cluster configuration has only one CRS framework resource and the component will prevent you from attempting to deploy more than one. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is crs-framework |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
| resource_dependencies |
The resource dependencies list |
Mandatory – default is rac-framework |
| resource_dependencies_offline_restart |
The resource offline restart dependencies list |
Mandatory – must be set to a scalable mount point or scalable device group resource. |
Top
RAC Proxy Resource Component (RAC_proxy_RS)
The RAC_proxy_RS component enables you to create the Sun Cluster RAC proxy resource that is used to proxy the monitoring, startup, and shutdown of Oracle RAC databases. This feature is only available on installations that run Sun Cluster 3.2 software. A Sun Cluster configuration can have multiple RAC proxy resources. Deploy this resource into a RAC proxy resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac-server-proxy-rs |
| rg_name |
The resource group name |
Mandatory – default is rac-server-proxy-rg |
| resource_dependencies |
The resource dependencies list |
Optional – default is rac-framework |
| resource_dependencies_offline_restart |
The resource offline restart dependencies list |
Optional – default is crs-framework, name_of_scalable_dg-resource or name_of_scalable_mnt_pt-resource |
| ORACLE_HOME |
The Oracle HOME of the RAC server under control |
Mandatory |
| CRS_HOME |
The Oracle HOME of the CRS server under control |
Mandatory |
| DB_NAME |
The Oracle database name being controlled |
Mandatory |
Top
RAC Solaris Volume Manager Resource Component (RAC_svm_RS)
The RAC_svm_RS component enables you to create the Sun Cluster RAC svm resource that is needed to drive the cluster Solaris Volume Manager reconfiguration and fencing process that is used to protect data that is held on the shared storage. Deploy this resource after you create the RAC_framework resource. A Sun Cluster configuration has only one RAC svm resource and the component will prevent you from attempting to deploy more than one. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac_svm |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
| resource_dependencies |
The resource dependencies list |
Mandatory – default is rac-framework |
Top
RAC Veritas Cluster Volume Manager Resource Component (RAC_cvm_RS)
The RAC_cvm_RS component enables you to create the Sun Cluster RAC cvm resource that is needed to drive the cluster Veritas Volume Manager reconfiguration and fencing process that is used to protect data held on the shared storage. Deploy this resource after you create the RAC_framework resource. A Sun Cluster configuration has only one RAC cvm resource and the component will prevent you from attempting to deploy more than one. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac_cvm |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
| resource_dependencies |
The resource dependencies list |
Mandatory – default is rac-framework |
| cvm_start_step_timeout |
The cvm start step timeout |
Mandatory – default is 120 |
| cvm_stop_step_timeout |
The cvm stop step timeout |
Mandatory – default is 40 |
| cvm_abort_step_timeout |
The cvm abort step timeout |
Mandatory – default is 40 |
| cvm_return_step_timeout |
The cvm return step timeout |
Mandatory – default is 40 |
| cvm_step1_timeout |
The cvm step 1 timeout |
Mandatory – default is 100 |
| cvm_step2_timeout |
The cvm step 2 timeout |
Mandatory – default is 100 |
| cvm_step3_timeout |
The cvm step 3 timeout |
Mandatory – default is 240 |
| cvm_step4_timeout |
The cvm step 4 timeout |
Mandatory – default is 320 |
| vxclust_port |
The vxclust port to use |
Mandatory – default is 5568 |
| vxclust_num_ports |
The number of vxclust ports to use |
Mandatory – default is 32 |
| vxkmsgd_port |
The vxclust port to use |
Mandatory – default is 5559 |
| vxconfigd_port |
The vxclust port to use |
Mandatory – default is 5560 |
Top
RAC Hardware RAID Resource Component (RAC_hwraid_RS)
The RAC_hwraid_RS component enables you to create the Sun Cluster RAC hwraid resource that is needed to drive the cluster hardware RAID reconfiguration and fencing process that is used to protect data held on the shared storage. Deploy this resource after you create the RAC_framework resource. A Sun Cluster configuration has only one RAC hwraid resource and the component will prevent you from attempting to deploy more than one. This resource is does not exist in Sun Cluster 3.2 software. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac_hwraid |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
| resource_dependencies |
The resource dependencies list |
Mandatory – default is rac-framework |
Top
RAC UDLM Resource Component (RAC_udlm_RS)
The RAC_udlm_RS component enables you to create the Sun Cluster RAC udlm resource that is needed to control the UNIX Distributed Lock Manager on SPARC platforms only. Deploy this resource after you create the RAC_framework resource. A Sun Cluster configuration has only one RAC udlm resource and the component will prevent you from attempting to deploy more than one. Deploy this resource into a RAC framework resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory – default is rac_udlm |
| rg_name |
The resource group name |
Mandatory – default is rac-framework-rg |
| resource_dependencies |
The resource dependencies list |
Mandatory – default is rac-framework |
| udlm_start_step_timeout |
The UDLM start step timeout |
Mandatory – default is 100 |
| udlm_abort_step_timeout |
The UDLM abort step timeout |
Mandatory – default is 100 |
| udlm_step1_timeout |
The UDLM step 1 timeout |
Mandatory – default is 100 |
| udlm_step2_timeout |
The UDLM step 2 timeout |
Mandatory – default is 100 |
| udlm_step3_timeout |
The UDLM step 3 timeout |
Mandatory – default is 100 |
| udlm_step4_timeout |
The UDLM step 4 timeout |
Mandatory – default is 100 |
| udlm_step5_timeout |
The UDLM step 5 timeout |
Mandatory – default is 100 |
| oracle_config_file |
The Oracle configuration file |
Mandatory – default is /etc/opt/SUNWcluster/conf/ udlm.conf |
| port |
The port to use |
Mandatory – default is 6000 |
| num_ports |
The number of ports |
Mandatory – default is 32 |
Top
Scalable Device Group Resource Component (Scalable_Device_Group_RS)
The Scalable_Device_Group_RS component enables you to create the Sun Cluster scalable device group resource that is used to monitor Solaris Volume Manager and Veritas Volume Manager shared device groups and volumes. This feature is only available on installations that run Sun Cluster 3.2 software. A Sun Cluster configuration can have multiple scalable device group resources. Deploy this resource into a scalable device group resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory |
| rg_name |
The resource group name |
Mandatory |
| resource_dependencies |
The resource dependency |
Mandatory |
| volume_manager |
svm or cvm |
Mandatory |
| DiskGroupName |
The disk group name |
Mandatory |
Top
Scalable Mount Point Resource Component (Scalable_Mount_Point_RS)
The Scalable_Mount_Point_RS component enables you to create the Sun Cluster scalable mount point resource that is used to monitor shared QFS and Network Attached Storage (NAS) mount points. This feature is only available on installations that run Sun Cluster 3.2 software. A Sun Cluster configuration can have multiple scalable mount point resources. Deploy this resource into a scalable mount point resource group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory |
| rg_name |
The resource group name |
Mandatory |
| resource_dependencies |
The resource dependency |
Optional |
| resource_dependencies_offline_restart |
The resource offline restart dependencies list |
Optional |
| MountPointDir |
The mount point directory |
Mandatory |
| FileSystemType |
The file system type s-qfs or nas |
Mandatory – default s-qfs |
| TargetFileSystem |
The target file system |
Mandatory |
Top
Shared QFS File System Resource Component (Shared_QFS_RS)
The Shared_QFS_RS component enables you to create the Sun Cluster shared QFS resource that is used to make the shared QFS metadata server highly available. A Sun Cluster configuration can have multiple shared QFS resources. Deploy this resource into a shared QFS group that is created or attached to by using the plug-in.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The resource name |
Mandatory |
| rg_name |
The resource group name |
Mandatory |
| QFSFileSystem |
The QFS file system mount point |
Mandatory |
Oracle Components
This section describes the following Oracle components:
Top
Oracle Clusterware Component (CRS)
The CRS component enables you to install or remove the Oracle Clusterware framework on one or more Sun Cluster nodes. A node can only run one instance of Oracle Clusterware. The component uses the host variables (where the host type is com.sun.oracle10g_RAC#OracleCRS_HT) of the plan targets.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installPath |
The base (ORACLE_HOME) directory for CRS |
Mandatory – default is /oracle/oracle/product/10.2.0/crs_1 |
| ORACLE_HOME_NAME |
The ORACLE_HOME_NAME for the ORACLE_HOME location |
Mandatory – default is OraCrs10g_home1 |
| ORACLE_BASE |
The ORACLE_BASE directory |
Mandatory – default is /oracle |
| COMPONENT_LANGUAGES |
The comma-separated list of languages to be installed |
Mandatory – default is en |
| ORACLE_CLUSTER_NAME |
The name of the CRS cluster |
Mandatory – default is crs |
| nodeList |
A comma-separated list of nodes that will run CRS, such as node1,node2,node3 |
Mandatory |
| ocrDevList |
A comma-separated list of two OCR disk devices or QFS files, or one if externally redundant. Must be 100 MB in size. |
Mandatory |
| vdskDevList |
A comma-separated list of three CRS voting disks or QFS files, or one if externally redundant. Must be 20 MB in size. |
Mandatory |
| mediaLocation |
Parent directory of the Oracle 10g RAC CRS release media that contains the runInstaller executable |
Mandatory |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Oracle RAC Component (RAC)
The RAC component enables you to install or remove the Oracle RAC software on one or more Sun Cluster nodes. The RAC software can be deployed more than once to any target node into different ORACLE_HOME locations.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installPath |
The ORACLE_HOME directory for RAC |
Mandatory – default is /oracle/oracle/product/10.2.0/db_1 |
| ORACLE_HOME_NAME |
The name for the ORACLE_HOME location |
Mandatory – default is OraDb10g_home1 |
| COMPONENT_LANGUAGES |
The comma-separated list of the languages to be installed |
Mandatory – default is en |
| install_type |
The installation type, EE or SE |
Mandatory – default is EE |
| nodeList |
The list of RAC cluster nodes |
Mandatory |
| mediaLocation |
Parent directory of the Oracle release media, which contains the runInstaller executable |
Mandatory |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Automatic Storage Management (ASM) Database Component (ASM)
The ASM component enables you to create, expand, shrink, or remove an Oracle Automatic Storage Management (ASM) database on one or more Sun Cluster nodes. Oracle permits only one ASM database on a cluster.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installName |
The database area disk group name |
Mandatory – default is my_asm_data_group |
| installPath |
The ORACLE_HOME with which to create the ASM database |
Mandatory – default is /oracle/oracle/product/10.2.0/db_1 |
| diskString |
The disk discovery path to be used by ASM |
Mandatory – default is /dev/did/rdsk/* |
| diskList |
A comma-separated list of disks for the database area disk group |
Mandatory |
| redundancy |
Must be set to HIGH, NORMAL, or EXTERNAL |
Mandatory – default is EXTERNAL |
| recoveryDiskList |
The comma-separated list of disks for the database area disk group |
Optional |
| recoveryGroupName |
Database area disk group name |
Optional – default is my_asm_recovery_group |
| recoveryGroupRedundancy |
Must be set to HIGH, NORMAL, or EXTERNAL |
Optional |
| emConfiguration |
Must be set to CENTRAL or NONE |
Mandatory – default is NONE |
| centralAgent |
Enterprise Manager central agent home (only used if emConfiguration is set to CENTRAL) |
Optional |
| nodeList |
The comma-separated list of nodes to create the ASM instances on |
Mandatory |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Oracle Database Stored on ASM Component (Database_on_ASM)
The Database_on_ASM component enables you to create, expand, shrink, or remove Oracle databases (or instances) on one or more Sun Cluster nodes where the storage is managed by Oracle's ASM database. You can deploy multiple ASM managed databases to the same Sun Cluster configuration, provided that the databases have different ORACLE_SID values.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installPath |
The ORACLE_HOME with which the database should be created |
Mandatory – default is /oracle/oracle/product/10.2.0/ db_1 |
| installName |
The Oracle SID |
Mandatory – default is orcl |
| GLOBAL_DB_NAME |
The global database name |
Mandatory – default is orcl.acme.com |
| templateName |
The name of the template to create the database from |
Mandatory |
| emConfiguration |
How Oracle Enterprise Manager will be configured. Valid entries are: CENTRAL, LOCAL, ALL, NOBACKUP , NOEMAIL, or NONE |
Mandatory |
| hostUserName |
The host user name for Enterprise Manager backup jobs (dependent option) |
Conditional |
| backupSchedule |
Daily backup schedule in the form of hh:mm (dependent option) |
Conditional |
| smtpServer |
Outgoing mail (SMTP) server for email notification (dependent option) |
Conditional |
| emailAddress |
Email address for email notifications (dependent option) |
Conditional |
| centralAgent |
Enterprise Manager central agent home (dependent option) |
Conditional |
| nodelist |
The comma-separated list of nodes to create the database on |
Mandatory |
| characterSet |
Character set for the database |
Mandatory |
| nationalCharacterSet |
National character set for the database |
Mandatory |
| initParams |
The comma-separated list of name=value pairs |
Optional |
| listeners |
List of listeners to configure the database with |
Optional |
| memoryPercentage |
The percentage of physical memory for Oracle to use |
Optional |
| diskGroupName |
The database area disk group name |
Mandatory |
| redundancy |
Set to HIGH, NORMAL, or EXTERNAL |
Mandatory |
| recoveryGroupName |
The database recovery area disk group name |
Optional |
| recoveryGroupRedundancy |
Set to HIGH, NORMAL, or EXTERNAL |
Optional unless recoveryGroupName is set |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Oracle Database Stored on Shared QFS File System Component (Database_on_QFS)
The Database_on_QFS component enables you to create, expand, shrink, or remove Oracle databases (or instances) on one or more Sun Cluster nodes where the data is stored on a shared QFS file system. You can deploy multiple databases, whose data is stored on shared QFS, to the same Sun Cluster configuration, provided that the databases have different ORACLE_SID values.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installPath |
The ORACLE_HOME with which the database should be created |
Mandatory – default is /oracle/oracle/product/10.2.0/ db_1 |
| installName |
The Oracle SID |
Mandatory – default is orcl |
| GLOBAL_DB_NAME |
The global database name |
Mandatory – default is orcl.acme.com |
| templateName |
The name of the template to create the database from |
Mandatory |
| emConfiguration |
How Oracle Enterprise Manager will be configured. Valid entries are: CENTRAL, LOCAL, ALL, NOBACKUP , NOEMAIL, or NONE |
Mandatory |
| hostUserName |
The host user name for Enterprise Manager backup jobs (dependent option) |
Conditional |
| backupSchedule |
Daily backup schedule in the form of hh:mm (dependent option) |
Conditional |
| smtpServer |
Outgoing mail (SMTP) server for email notification (dependent option) |
Conditional |
| emailAddress |
Email address for email notifications (dependent option) |
Conditional |
| centralAgent |
Enterprise Manager central agent home (dependent option) |
Conditional |
| nodelist |
The comma-separated list of nodes to create the database on |
Mandatory |
| characterSet |
Character set for the database |
Mandatory |
| nationalCharacterSet |
National character set for the database |
Mandatory |
| initParams |
The comma-separated list of name=value pairs |
Optional |
| listeners |
List of listeners to configure the database with |
Optional |
| memoryPercentage |
The percentage of physical memory for Oracle to use |
Optional |
| datafileDestination |
The destination directory for all database files |
Mandatory |
| recoveryAreaDestination |
The destination directory for all recovery files |
Optional |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
Top
Oracle Database Stored on raw Disk Component (Database_on_RAW)
The Database_on_RAW component enables you to create, shrink, or remove Oracle databases (or instances) on one or more Sun Cluster nodes where the data is stored on raw disk (either hardware RAID LUNs or shared device groups). You can deploy multiple databases, whose data is stored on raw devices, to the same Sun Cluster configuration, provided that the databases have different ORACLE_SID values and use different raw devices.
The component has the following configurable component variables.
| Name |
Description |
Notes |
| installPath |
The ORACLE_HOME with which the database should be created |
Mandatory – default is /oracle/oracle/product/10.2.0/ db_1 |
| installName |
The Oracle SID |
Mandatory – default is orcl |
| GLOBAL_DB_NAME |
The global database name |
Mandatory – default is orcl.acme.com |
| templateName |
The name of the template to create the database from |
Mandatory |
| emConfiguration |
How Oracle Enterprise Manager will be configured. Valid entries are: CENTRAL, LOCAL, ALL, NOBACKUP , NOEMAIL, or NONE |
Mandatory |
| hostUserName |
The host user name for Enterprise Manager backup jobs (dependent option) |
Conditional |
| backupSchedule |
Daily backup schedule in the form of hh:mm (dependent option) |
Conditional |
| smtpServer |
Outgoing mail (SMTP) server for email notification (dependent option) |
Conditional |
| emailAddress |
Email address for email notifications (dependent option) |
Conditional |
| centralAgent |
Enterprise Manager central agent home (dependent option) |
Conditional |
| nodelist |
The comma-separated list of nodes to create the database on |
Mandatory |
| characterSet |
Character set for the database |
Mandatory |
| nationalCharacterSet |
National character set for the database |
Mandatory |
| initParams |
The comma-separated list of name=value pairs |
Optional |
| listeners |
List of listeners to configure the database with |
Optional |
| memoryPercentage |
The percentage of physical memory for Oracle to use |
Optional |
| datafileNames |
Full path name to a file that contains the raw disk devices to be used for the database creation. Entries will be name=value pairs |
Mandatory |
The nodes that are specified in the nodeList parameter should be the physical hostnames of the target systems, that is, the result that is returned from the /usr/bin/uname -n command.
The raw-disk device map file can be checked into N1 SPS separately and distributed to the target machines, though this feature is not offered by the plug-in itself.
Example: Sample Raw Disk Map File
% cat /tmp/raw_disk_map.txt
# Sample raw disk map file for a four node cluster
sysaux=/dev/md/oraset/rdsk/d106
example=/dev/md/oraset/rdsk/d107
users=/dev/md/oraset/rdsk/d108
temp=/dev/md/oraset/rdsk/d109
undotbs1=/dev/md/oraset/rdsk/d110
undotbs2=/dev/md/oraset/rdsk/d111
undotbs3=/dev/md/oraset/rdsk/d112
undotbs4=/dev/md/oraset/rdsk/d113
redo1_1=/dev/md/oraset/rdsk/d114
redo1_2=/dev/md/oraset/rdsk/d115
redo2_1=/dev/md/oraset/rdsk/d116
redo2_2=/dev/md/oraset/rdsk/d117
redo3_1=/dev/md/oraset/rdsk/d118
redo3_2=/dev/md/oraset/rdsk/d119
redo4_1=/dev/md/oraset/rdsk/d120
redo4_2=/dev/md/oraset/rdsk/d121
control1=/dev/md/oraset/rdsk/d122
control2=/dev/md/oraset/rdsk/d123
spfile=/dev/md/oraset/rdsk/d124
pwdfile=/dev/md/oraset/rdsk/d125
system=/dev/md/oraset/rdsk/d126
|
Sign up or Log in to add a comment or watch this page.