| eG Monitoring |
|---|
|
Measures reported by OraRacLsStTest Oracle RAC 11g release 2 introduces the Single Client Access Name (SCAN), which provides a single name for clients to access Oracle Databases running in a cluster and simplify the database connection strings that an Oracle Client uses to connect. Without SCAN, till Oracle Clusterware 11g release 1, every client and server that participated in the RAC environment needed to have a tnsnames.ora file, which contains a single entry for every node in the cluster. This meant that, whenever the cluster configuration changes - say, a node is added/removed from the cluster - the tnsnames.ora file of every client had to be updated with this change. SCAN on the other hand, completely eliminates the need for these edits! In short, with SCAN, clients can access any database in the cluster, independent of the number of databases or servers running in the cluster, or the server(s) on which the requested database is actually active. The RAC components that play a crucial role in SCAN are as follows:
SCAN listeners can run on any node in the cluster. Typically, each cluster will have a minimum of 1 and a maximum of 3 SCAN listeners, regardless of the number of nodes in the cluster. Three SCAN listeners are recommended considering load balancing and high availability requirements. A SCAN listener can be started only when the SCAN VIPs are online. The root.sh script automatically configures the SCAN VIPs and SCAN listeners. When the root.sh is run on the first node, all 3 scan IPs are plugged on to the first node's public interface. The SCAN/VIP combination will failover to another node in the cluster, if the current node fails. Likewise, SCAN also load balances across the instances providing a service, by selecting and routing requests to the least loaded node. Also, if a single SCAN listener IP stops running on a node, connection requests will automatically failover to another SCAN listener IP address on that node. This means that, if all SCAN listeners in a cluster are down or are not enabled, then users will not be able to connect to the database service. To avoid this, administrators must track the status of each SCAN listener in an RAC cluster, promptly identify the listener that is down, and quickly start it. This can be achieved using the OraRacLsStTest. This test auto-discovers the SCAN listeners in a target cluster, and reports the status of each listener. In the process, the test alerts administrators to listeners that are not enabled or are not running. Detailed diagnostics point administrators to the exact node on which the failure has occurred, thereby easing troubleshooting. Outputs of the test: One set of results for every SCAN listener. The measures made by this test are as follows:
|