Shulou(Shulou.com)05/31 Report--

How to carry out ORACLE 11GR2 RAC network planning analysis, many novices are not very clear about this, in order to help you solve this problem, the following editor will explain for you in detail, people with this need can come to learn, I hope you can gain something.

In ORACLE 11gR2, the installation of RAC has changed significantly. In the era of 10g and 11gR1, the step to install RAC is to install CRS, and then install DB, while in the era of 11gR2, crs and asm are integrated together, collectively known as GRID, you must first install GRID before you can continue to install DB, otherwise, you will not have a chance with 11gR2's RAC, hehe. Let's talk about the network planning of ORACLE 11g R2 RAC. Starting from ORACLE 11.2, there are special requirements for network IP addresses and the addition of SCAN IP, so at least four IP addresses are required from 11.2.

Understand the partition planning in shared storage. Look at the figure below.

Suppose there are 4 network cards on a server: eth0,eth2,eth3,eth4.

You can bind eth0 and eth3 into bond0. As the public-ip of RAC, it provides external communication.

Then the eth2 and eth4 are bound into bond1, which serves as the private-ip of RAC to provide internal heartbeat communication.

The HBA card on the server communicates with the back-end storage through the fiber switch.

Let's take a look at the contents of the / etc/hosts file

10.18.12.140 S1P32 # RAC1 Public

10.18.12.142 S1P32-vip # RAC1 VIP

22.22.22.140 S1P32-priv # RAC1 Private

10.18.12.141 S1P33 # RAC2 Public

10.18.12.143 S1P33-vip # RAC2 VIP

22.22.22.141 S1P33-priv # RAC2 Private

10.18.12.144 catdb-scan # RAC SCAN

Let's talk about PUBLIC and VIP first.

Each node in the Oracle RAC has a virtual IP, referred to as VIP, which is on the same network segment as the public network PUBLIC IP. Vip is attached to the public interface.

The main difference between VIP and PUBLIC IP is that VIP is floating while PUBLIC IP is fixed. When all nodes are running normally, the VIP of each node will be assigned to the public NIC. According to the ifconfig view under linux, there are 2 IP addresses on the public Nic. If a node goes down, the VIP of that node will be transferred to the running node. That is, the public NIC of the surviving node will have three IP addresses on the network card.

From the figure, it can be seen that the PUBLIC IP address is a public address bound to a dual network card, and external users access it through switch S1.

2. Look at the Private in RAC.

The Private private IP in RAC is used for heartbeat synchronization, which can be ignored at the user level. To put it simply, this Ip is used to ensure that the two servers synchronize data, which belongs to the communication between RAC. Priv and public should not belong to the same interface.

Another question that everyone asks: when doing RAC, can the network card used for the interconnection between two node be connected by cross cable?

The RAC document on Metalink recommends that the switch be used as the connection of the internal network card instead of the crossover cable, to avoid removing the binding protocol because the network card detects the link failure state due to the shutdown or restart of the connected node. Causes the cache merge network to become unavailable.

From the figure, the internal communication between the two nodes interacts through switch S2.

3. SCAN IP in RAC

In 11gR2, SCAN IP appears as a new IP. Scan ip is actually a new connection layer added by oracle between the client and the database. When there is client access, it connects to the SCAN IP LISTENER, and when SCAN IP LISTENER receives the connection request, it will forward the client connection request to the corresponding VIP LISTENER on the instance according to the LBA algorithm, thus completing the connection process between the client and the server. The simplification is as follows:

Client-> scan listener-> local listener-> local instance

You can also think of scan as a virtual hostname that corresponds to the entire RAC cluster. Client hosts only need to use this scan name to access any node of the database cluster. Of course, the nodes visited are random. Oracle strongly recommends that you resolve SCAN through DNS Server's round robin mode configuration to achieve load balancing (that is, rotate the IP address corresponding to the SCAN connection). This is somewhat similar to the principle of load balancing through vip and listener loadbalance configuration.

Oracle RAC itself is relatively complex, in the installation and management may encounter a variety of problems, involving OS, RDBMS, Cluster software and network, host, storage and other hardware, in order to avoid unnecessary problems, before the installation of a clear network planning for the entire architecture is very important, would like to share with you to learn.

Is it helpful for you to read the above content? If you want to know more about the relevant knowledge or read more related articles, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.