The introduction of SCAN have caused a lot of confusion among network administrators, architects and DBAs, let's see if we can keep this to basis to make it less confusing.
Why do we need this?
It provides a single name for clients to access Oracle Databases, this is specially true for database clusters, the reason is because you do not have to update later on the client information if you change the cluster configuration.
SCAN gets configured when you install Oracle Grid Infrastructure (New name for Clusterware and ASM together), During the installation you will be asked for a SCAN name, ideally you want to define the SCAN in your DNS, with this method your network administrator needs to create a virtual hostname that resolves to 3 IP addresses using a round robin algorithm. The reason behind why 3 IP Addresses is related to load balancing and HA requirements regardless of the number of servers in your cluster. Each IP must be on the same subnet as the public network in your database cluster.
NOTE: The 3 IPs above should not be assigned to a network interface, Oracle Clusterware will take care of it.
Your DNS entry should looks like
blog-scan-oraprofessionals.com IN A 192.168.1.102
IN A 192.168.1.103
IN A 192.168.1.104
How it works?
Clients connecting to the database using the SCAN will receive 3 IP, the client will try to connect using the list of IPs it received, if the client receives an error , it will try the other addresses before it returns an error. When the SCAN listener received a connection request, it checks for the least loaded instance providing the requested service, then it re-directs the connection to the local listener on the node that have the service available.
Now as you can see by the explanation above, your client connection information won't be affected by the configuration of your cluster.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment