Types of IP In Oracle RAC and Its Usage


When we talk about the oracle RAC cluster database then it means we’re talking about the several independent servers which are physically different but logically or at a cluster level they’re working as a single database machine, to connect multiple physical servers as a single cluster server we need multiple types of IP and each of it has its own usage.

 Public IP

It belongs to the client network in the network architecture of cluster database system, it is the IP that is assigned during the installation of the GI. The public IP address is the primary address for the member of the nodes and it requires to connect with the internet or we can say to connect with the database from public network.

Private IP

 Private IP is use for inter connectivity between the nodes to transfer the data and also to transfer the heartbeat between each and every node. The clusterware uses the TCP protocol for intercommunication between the nodes or sending the heartbeat information and at the same time the clusterware uses the UDP protocol to transfer the database between the nodes to maintain the cache coherency.

                                 Interconnect Failure in Private IP

The CTSS background process monitor the heart beat of every node in the cluster so that they’ll not create an independent cluster inside the same cluster which may lead the cluster to data corruption or data inconsistency and for that the CTSS checks two type of heart beat from every node.

v Disk heartbeat: each node in the cluster stores their heart beat information in a voting disk on a periodical basis

v Network Heartbeat: each node in the cluster send their heart beat information to each and every node on a periodical basis.

How To Check Private Interconnect for Oracle Clusterware

[oracle@oratest ~]$ olsnodes -l -p

oratest        192.168.24.30

                                  VIP

Each and every node into the cluster has its own VIP address and it helps the user connection to send the user’s request to the local listener of the node since it’s used by the end user so the VIP address must run in the same subnet of the public IP address. It can be registered either using GNS, DNS or in a hosts file.

                             Scan IP

Scan IP is a part of client network connect and the end users basically connects with the database using the scan name which eventually fetch the IP address from DNS server and after that it also help the user’s connection to re-route to the least loaded instance on the basis of load balancing advisory maintained by LREG background process.

                v The scan listener created for each scan VIP

v A dependency for the scan listener is configured for the SCAN VIP

v The scan must run in a round robin fashion to perform load balancing

          Benefit Of Using A Scan

(1)             The main benefit of using a scan that the client can connect with the database using a single scan name , prior to the introduction of the scan the client need to update the each VIP and host information in their tns file and whenever the nodes join or delete from the cluster then they also have to update the VIP information accordingly.

(2)            Prior to the introduction of scan if the node goes down then the client has to wait for the timeout however after the introduction of scan when the nodes go down the CRS will immediately come to know and failover the vip resource in the surviving instance.




 

Post a Comment

Previous Post Next Post