| To make network data constantly available through your newly created cluster, you need to create and configure shared cluster pools and volumes. Shared volumes can be any of the traditional Linux filesystems (Reisfer, EXT2/3, and so on) or NSS. Shared NSS volumes must be created on shared NSS pools. NSS pools can be designated as sharable either during or after creation. This is possible through both iManager and the command-line nssmu utility. To create a cluster-enabled NSS pool during pool creation via iManager, complete the following steps: 1. | From within iManager, select the Pools link, found under the Storage category in the left navigation frame.
| 2. | In the right pane, use the Object Selector or Object Browser tools to locate your cluster object. After you select the cluster object, the current page will automatically reload.
| 3. | Select New to create a new pool.
| 4. | Enter the name of the NSS pool, and click Next.
NOTE Because periods are used as delimiters in fully qualified names in eDirectory, periods cannot be used in the naming of your NSS pool. If spaces are used in the name, eDirectory will convert those spaces to underscores. | 5. | Check the box next to the device on which the NSS pool should be created. Also enter the size of the new NSS pool. If desired, select the Mount on Creation check box to make the pool available after creation.
| 6. | Select Finish to create the new NSS pool. For more information on NSS pools, see Chapter 11.
| 7. | You can either select to cluster-enable the pool now or later. You must cluster-enable a pool in order for it to fail over during a failure. If you choose to cluster-enable the storage pool now, you have to provide the following information:
- Virtual Server Name Change the name of the default Virtual Server object. When you cluster-enable a pool, the Virtual Server object is named by combining the Cluster object name and the Pool object name. For example: QuillsCluster_SharePool_Server.
- CIFS Server Name If you select CIFS as an advertisement protocol, specify a server name that CIFS clients will see for this storage when browsing the network.
- IP Address Each cluster-enabled pool requires its own IP address. This IP address is used to provide access and failover capability to the pool. It is assigned to the storage pool and associated with volume within the pool. All volumes in the storage pool share the same IP address.
- Advertising Protocols Specify how you want the shared storage pool to advertise its existence to clients. AFP is used by Macintosh clients, CIFS is used by Microsoft Windows, and NCP is used by the Novell client.
|
Cluster-Enabling a Volume After Pool Creation When you have created the NSS pool, you can add sharable NSS volumes. Shared volumes can be created with either iManager or nssmu. The following steps can be used to create a shared volume using iManager: 1. | From within iManager, select the Volumes link, found under the Storage category in the left navigation frame.
| 2. | In the right pane, use the Object Selector or Object Browser tools to locate your cluster object. After you select the cluster object, the current page will automatically reload.
| 3. | Select New to create a new volume.
| 4. | Enter a name for the new volume and click Next.
| 5. | Check the box next to the pool on which the NSS volume should be created. Also enter the size of the new NSS volume.
| 6. | Select the desired NSS attributes for the volume. Also, designate a mount point and select whether to mount and activate the volume on creation. Click Finish to complete the volume creation.
|
The whole point in creating a cluster is to provide constant access to network resources. Because one of the principal resources on a network is data, cluster-enabling a storage pool will likely be one of the things you do first. NOTE To cluster-enable a non-NSS volume, create the Linux partition on the shared storage system as you normally would. After creation, create a cluster resource with the proper load and unload script to provide access to the partition.
Cluster-Enabling a Storage Pool After Creation If you chose not to cluster-enable a storage pool when it was initially created, you can do so after the fact from iManager. Make sure you deactivate the pool and dismount the volume(s) before cluster-enabling them. To cluster-enable an existing pool, complete the following steps: 1. | From within iManager, select the Cluster Options link, found under the Clusters category in the left navigation frame.
| 2. | In the right pane, use the Object Selector or Object Browser tools to locate your cluster object. After you select the cluster object, the current page will automatically reload.
| 3. | Select New from the Cluster Objects window.
| 4. | Specify Pool as the new object type, and then click Next.
| 5. | Enter the name of the pool you would like to cluster-enable, or browse and select one. After a pool has been selected, complete the following information and click Next:
- Virtual Server Name Enter the name of the Virtual Server object. When you cluster-enable a pool during pool creation, the Virtual Server object is named by combining the Cluster object name and the Pool object name, for example: QuillsCluster_SharePool_Server.
- CIFS Server Name If you select CIFS as an advertisement protocol, specify a server name that CIFS clients will see for this storage when browsing the network.
- IP Address Each cluster-enabled pool requires its own IP address. This IP address is used to provide access and failover capability to the pool. It is assigned to the storage pool and associated with volume within the pool. All volumes in the storage pool share the same IP address.
- Advertising Protocols Specify how you want the shared storage pool to advertise its existence to clients. AFP is used by Macintosh clients, CIFS is used by Microsoft Windows, and NCP is used by the Novell client.
- Online Resource After Create Select this check box to automatically start the volume resource immediately after object creation.
- Define Additional Properties Select this check box to continue cluster configuration of the pool following object creation.
|
You can delete cluster-enabled volumes and pools in the same way that you delete standard NSS volumes and pools. When a cluster-enabled resource is deleted, NCS will automatically modify its load scripts to remove that resource. Modifying Node Assignment for a Storage Pool When you cluster-enable a storage pool, all nodes in the cluster are automatically assigned to the pool. The order of assignment is the order in which the nodes appear in the list. To assign or unassign nodes, or to change the failover order, complete the following steps: 1. | From within iManager, select the Cluster Options link, found under the Clusters category in the left navigation frame.
| 2. | In the right pane, use the Object Selector or Object Browser tools to locate your cluster object. After you select the cluster object, the current page will automatically reload.
| 3. | Select the check box next to the storage pool whose start and failover modes you are changing. When the check box is selected, click the Properties link in the Cluster Objects window to access the properties of the storage pool.
| 4. | Select the General tab.
| 5. | From the list of unassigned nodes, select the server you would like to assign the storage pool to. Click the right arrow to move the server to the Assigned Nodes list. This process should be repeated for all servers that should be assigned to the resource.
| 6. | Click the up and down arrows to adjust the failover order of the servers assigned to the storage pool. Click Apply or Finish to save your changes.
|
The first server in the Nodes list will be the preferred node for the cluster-enabled storage pool. Failover will occur sequentially down the list. Configuring Storage Pool Policies After a storage pool has been cluster-enabled, you can configure the start, failover, and failback parameters. To do this, complete the following steps: 1. | From within iManager, select the Cluster Options link, found under the Clusters category in the left navigation frame.
| 2. | In the right pane, use the Object Selector or Object Browser tools to locate your cluster object. After you select the cluster object, the current page will automatically reload.
| 3. | Select the check box next to the storage pool whose start, failover, or failback modes you would like to edit. Then click the Properties link in the Cluster Objects window.
| 4. | Select the General tab, and configure the following options:
- Resource Follows Master Check this box if you want the resource to run only on the master node in the cluster. If the master node fails, the resource will fail over to whichever node becomes the master.
- Ignore Quorum Check this box if you don't want the cluster-wide timeout period and node number limit enforced. This makes sure the resource is launched immediately as soon as any server in the Assigned Nodes list is brought online. You can modify the quorum values from the Cluster Configuration page in iManager.
| 5. | Continue with the storage pool configuration by adjusting the following options:
- Start When this option is set to Auto, the resource will start automatically whenever the cluster is brought online. When this option is set to Manual, you must start the device after the cluster comes online. The default is Auto.
- Failover When this option is set to Auto, the resource will automatically move to the next server in the Assigned Nodes list if the node it is currently running on fails. When this option is set to Manual, you will intervene after a failure and reassign the resource to a functioning node. The default is Auto.
- Failback When this is set to Auto, the cluster resource will migrate back to its preferred node when it comes back online. The preferred node is the first node listed in its Assigned Nodes table. When this option is set to Manual, the cluster resource will not fail back until you allow it to happen. When this option is set to Disable, the cluster resource will not fail back to its most preferred node when the most preferred node rejoins the cluster. The default is Disable.
| 6. | Click Apply to save the configuration changes.
|
Adjusting these settings gives you granular control over the behavior of your clustered storage pool. |
