Redhat 5.8 ORACLE 11gR2 RAC installation Documentation 2-grid installation

3. Installing the grid

3.1. Install RPM Package

Two nodes are to be installed, take node 1 as an example

[Email protected] yum.repos.d]# yum install compat-libstdc++-33 ksh gcc gcc-c++ libgomp elfutils-libelf-devel glibc- Devel glibc-headers libaio-devel libstdc++-devel sysstat UnixODBC unixodbc-devel–y ................. complete! ?

Remember to install the RPM package in the same way as the second node. The installation of Yum is also very convenient, if you find RPM–IVH installed separately, all kinds of dependency relationship error.

3.2, CVU detection before installation

3.2.1 Upload and unzip the installation package B

Can be mounted through the CD-ROM media or directly upload the downloaded compressed package, a total of three compressed packets, the first two are the database package, the third is a grid of compressed packets.

-r-xr-xr-x 1 root root 1358454646 2015-06-17 18:13 P10404530_112030_linux-x86-64_1of7.zip -r-xr-xr-x 1 root root 1142195302 2015-06-17 18:18 p10404530_112030_linux-x86-64_2of7.zip P style= "text-align:justify" >-r-xr-xr-x 1 root root 979195792 2015-06-17 18:22 p10404530_112030_linux-x86-64_3of7.zip

See two directories when the decompression is complete

/tr>

[[email protected" src]# ls-lrt total 3402840 drwxr-xr-x 8 root root 4096 2011-09-22 16:57 grid drwxr-xr-x 8 root root 4096 2011-09-22 16:57 database

3.2.2 Installing the CVU RPM package

Two nodes are to be installed

Node 1:

[email protected] grid]# cd/usr/local/src/grid/rpm/ [[email protected] rpm]# ls cvuqdisk-1.0.9-1.rpm [Email protected] rpm]# rpm-ivh cvuqdisk-1.0.9-1.rpm # installation Preparing ... ########################################### [100%] Using the default group Oinstall to the install package 1:cvuqdisk ########################################### [100%] [email protected] rpm]# SCP cvuqdisk-1.0.9-1.rpm orcl2:/usr/local/src/#上传到另外一个节点 [email protected]' s password: cvuqdisk-1.0.9-1.rpm 100% 8551 8.4kb/s 00:00

Node 2:

[email protected] ~]# cd/usr/local/src [email protected] src]# rpm-ivh cvuqdisk-1.0.9-1.rpm #安装 Preparing ... ########################################### [100%] Using the default group Oinstall to the install package 1:cvuqdisk ########################################### [100%]

3.2.3 Running CVU detection scripts

operate on Node 1 only

[email protected] grid]$/runcluvfy.sh stage-pre crsinst-n orcl1.demo.com,orcl2.demo.com-fixup-verbose ; /tmp/result.txt Open a terminal window to track the test results. [Email protected] grid]$ tail-f/tmp/result.txt #跟踪检测结果, if you have faild to adjust after the prompt to re-detect, repeated several times until the following results appear Pre-Check for Cluster Services Setup was successful.?

3.3. Installing the Grid

Installing the grid requires only the operation on node 1

Since my side is a former production server, so I do not configure the desktop environment, in order to facilitate my use of Xmanager software installed here.

3.3.1 Installation Xmanager

输入序列号081129-116771-999614

3.3.2 Configuring user profiles to increase call Xmanager parameters

Grid users:

[email protected] ~]$ vi. bash_profile #增加以下两行 Export display=172.16.40.19:0.0 #主要是这行 Stty Erase ^h #该行是sqlplus命令行下退格键删除

Oracle Users:

[email protected] ~]$ vi. bash_profile #增加以下两行 Export display=172.16.40.19:0.0 #主要是这行 Stty Erase ^h After the configuration is complete, remember that source. Bash_profile Effective

IP is the wireless IP address of the local client such as My Notebook

3.3.3 Installing the Grid

On the Windows side, double-click Open Xmanager Passive to receive remote graphics

[email protected] grid]$ ./runinstaller Starting Oracle Universal Installer ... ? Checking Temp space:must be greater than MB. Actual 479717 MB Passed Checking swap space:must be greater than MB. Actual 31743 MB Passed Checking Monitor:must is configured to display at least-colors. Actual 16777216 Passed Preparing to launch Oracle Universal Installer from/tmp/orainstall2015-06-18_04-07-47pm. Please wait. Executing the Runinstaller installation script lifting the installation interface

?

Choose Skip Software Updates,next

Select the first item, install for a cluster, Next

Select Second, advanced installation, Next

Add Simplified Chinese support, Next

Set cluster name and scan name, cancel check Configure Gns,next

At this point the installer does not automatically recognize Node2 information and needs to add it manually.

Add Good Node2,next

Ahab, error, this mistake is two nodes ssh mutual trust caused, in fact, oracle11g RAC can configure their own SSH trust, do not manually configured, but outside the front has done a manual configuration, so go to check, found 2 grid user under SSH orcl1 date Still need to enter Yes, after inspection found that I do not have a config, this place leaks, the specific steps can refer to the previous SSH trust step, is to create a config file in the. ssh directory, add stricthostkeychecking no this line, error can be.

Over and over again.

Set which NIC is a public or private network, set it up, Next

Choose Asm,next

, create ASM Disk groups for ocr&&voting disks, redundant mode selection Normal,next

?

The following are the explanations for OCR and Votedisk:

Oracle Cluster Registry (OCR) is used to solve the problem of amnesia in clusters.

Forgetfulness is due to the fact that a node has updated the contents of OCR, while other nodes in the cluster are in the shutdown, maintenance, or restart phase, and the OCR master process is too late to update its information to these exception node caches, resulting in inconsistencies.

For example, a command to add an OCR image is issued at the a node, at which point the B node is in the restart phase. After the reboot A has been updated, and at this point B does not know that the OCR has added a new mirror disk, this will cause the configuration is lost, that is, the so-called "amnesia."

?

Votedisk is a problem that is used to solve a brain fissure (split Brain).

? In the cluster, the nodes are aware of each other's health state through a mechanism (heartbeat) to ensure that the nodes coordinate their work. Assuming that only the "heartbeat" problem, the nodes are still working, at this point, each node is considered to be the other nodes down, oneself is the entire cluster environment "The only surviving", oneself should get the whole cluster "control." In a clustered environment, storage devices are shared, which means data disaster, which is "brain crack."

?

Three redundant modes of ASM

??????? High redundancy: ASM uses 3 copies of mirrored storage to improve performance and data security, requiring a minimum of three disks (three failure group), and valid disk space is 1/3 of the sum of all disk device sizes, although the redundancy level is high, but the hardware is the most expensive.

General redundancy (normal): ASM uses 2 copies of mirrored storage to improve performance and data security, requiring a minimum of two disks (two failure group), and valid disk space is 1/2 of the sum of all disk device sizes, typically with regular redundancy.

External redundancy (External): ASM does not mirror the database files, can be mirrored by a RAID disk, the disk is the least, and the effective disk space is the sum of the sizes of all disk devices.

Choose to use the same ASM password, password strength requirements must have letter case and number combination, otherwise there will be hints, (can be ignored) in order to facilitate memory, I set to Manager,next

Password does not meet complexity requirements, click Yes to ignore it. It is recommended that the production environment set complex passwords.

Choose not to use Ipmi,next

By default, Next

By default, Next

By default, Next

?

Wait for pre-installation detection

Did not find an error, click Install, if there is an error, according to the tips of troubleshooting.

Wait for the installation process for a long time

Execute the two scripts on the Node1,node2, respectively.

?

After the two-node script is executed successfully, click OK to continue the installation

Point Close,grid installation is complete.

?

Redhat 5.8 ORACLE 11gR2 RAC installation Documentation 2-grid installation