Basic Environment
Use two HP DL580 servers as RAC nodes, storage using IBM V7000. The specific environment is as follows:
Equipment |
Use |
IP Address |
Disk space |
HP DL580 |
RAC Node 01 |
rac01-pub:116.1.1.57 rac01-priv:4.4.4.35 rac01-vip:116.1.1.59 rac02-pub:16.1.1.58 rac02-priv:4.4.4.36 rac02-vip:116.1.1.60 rac-scan:116.1.1.61 |
300G |
HP DL580 |
RAC Node 02 |
300G |
IBM V7000 |
Shared storage |
|
ocr:30g fra:450g data1:600g data2:600g data3:600g |
The previous network configuration (two sets of dual-NIC bindings, respectively, the pub IP and Priv IP), the shared storage connection, and the Yum source configuration have been completed. Oracle version: 11.2.0.4, which stores multipath software as a multipath with Redhat.
implementation Stepspreparatory workhosts and host name configuration (all nodes)
To edit the/etc/sysconfig/network file configuration hostname (restart effective):
#所有节点均执行
[Email protected]:~]#cat /etc/sysconfig/network NETWORKING=yeshostname= Cheastrac01[[email protected]:~]#cat /etc/sysconfig/networknetworking=yeshostname =CHEASTRAC02
Edit/etc/hosts Configuration hosts:
#所有节点均执行
[[Email protected]:~]$Cat/etc/hosts127.0.0.1localhost localhost.localdomain localhost4 localhost4.localdomain4::1localhost localhost.localdomain localhost6 localhost6.localdomain6#rac116.1.1.57CHEASTRAC01116.1.1.58CHEASTRAC02116.1.1.59Cheastrac01-VIP116.1.1.60cheastrac02-VIP4.4.4.35Cheastrac01-Priv4.4.4.36cheastrac02-Priv116.1.1.61Cheastrac-scan
firewall, SELinux, and NTP disabled (all nodes)
To turn off the server firewall:
All nodes are executed
[[Email protected]:~] $service iptables stop [[email protected]:~] $chkconfig iptables off
Disable SELinux:
#所有节点均执行
[Email protected]:~]$sed's/selinux=enforcing/selinux=disabled/g' /etc/ selinux/Config
Disable NTP and delete its profile: (After disabling NTP and deleting the configuration file, the RAC's grid installation automatically uses its own CTSS for time management)
#所有节点均执行
[[Email protected]:~] $service ntpd stop [email protected]:~] $chkconfig ntpd off[[email protected]:~]$mv /etc/ntp.conf/etc/ Ntp.conf.bak
Resource Limit configuration (all nodes)
Configure resource limits for grids and Oracle.
Edit/etc/sysctl.conf and add the following:
#所有节点均执行
[[Email protected]:~]$Cat/etc/sysctl.conf. #oracle racfs .file-max =6815744Kernel.sem= - 32000 - -Kernel.shmmni=4096Kernel.shmall=1073741824Kernel.shmmax=96636764160#依当前内存而定, usually 60%- the%about Net.core.rmem_default=262144Net.core.rmem_max=4194304Net.core.wmem_default=262144Net.core.wmem_max=1048576Fs.aio-MAX-NR =1048576Net.ipv4.ip_local_port_range=9000 65500
[11:10 [email protected]:~] $sysctl-P #令配置生效
Edit/etc/security/limits.conf and add the following:
#所有节点均执行
[[Email protected]:~]$Cat/etc/security/limits.conf. #oracle Racgrid Soft Nofile1024x768grid Hard Nofile65536Grid Soft Nproc2047grid Hard Nproc16384Grid Soft Stack10240grid Hard Stack32768Oracle Soft Nofile1024x768Oracle Hard Nofile65536Oracle Soft Nproc2047Oracle Hard Nproc16384Oracle Soft Stack10240Oracle Hard Stack32768
Edit/etc/pam.d/login and add the following:
#所有节点均执行
[Email protected]:~] $echo "session required pam_limits.so" >>/etc/pam.d/login
Adjust TMPFS size
If the memory is large, the SGA target cannot exceed the TMPFS size, so you need to adjust the TMPFS to a value larger than the SGA target. Edit/etc/fstab, modify the TMPFS line:
#所有节点均执行
Tmpfs /dev/shm tmpfs defaults,size=102400m 00 # If the server memory 128GB,SGA target is 90GB, you can set TMPFS to 100GB
Re-mount and see if the TMPFS setting is in effect:
#所有节点均执行
[Email protected]:~] $mount-o remount/dev/shm [[email protected]:~] $DF-H
User and group configuration (all nodes)
Create a new grid and Oracle-related users and groups.
#所有节点均执行 and ensure that the same group and user ID for all nodes
[Email protected]:~] $groupadd-G4321Oinstall [[email protected]:~] $groupadd-G4322Dba[[email protected]:~] $groupadd-G4323Oper[[email protected]:~] $groupadd-G4324Asmdba[[email protected]:~] $groupadd-G4325Asmadmin
[Email protected]:~] $groupadd-G4326Asmoper
[email protected]:~] $useradd-U4321-G Oinstall-G asmadmin,asmdba,asmoper,dba-d/home/grid-s/bin/bash-c "grid Infrastructure Owner" grid
[[Email protected]:~] $useradd-u4322-G oinstall-g asmadmin,asmdba,dba,oper-d/home/oracle-s/bin/bash-c "Oracle software Owner" Oracle
[[Email protected]:~] $echo "Grid" | passwd--stdin Grid #为grid和oracle用户设置密码
[[Email protected]:~] $echo "Oracle" | passwd--stdin Oracle
New grid and Oracle installation directory
#所有节点均执行
[Email protected]:~] $mkdir-P/u01/grid/product/grid11g [[email protected]:~] $mkdir-P/u01/ oracle/product/db11g[[email protected]:~] $chown-R grid:oinstall/U01[[email protected]:~] $chown- R oracle:oinstall/u01/oracle[[email protected]:775 /u01
Configuring Grid and Oracle User environment variables
grid environment variables: Edit the/home/grid/.bash_profile file and add the following:
#所有节点均执行
[[email protected]:~] $cat/home/grid/.bash_profile Export Pathexport TMP =/tmpexport TMPDIR =$ Tmpexport oracle_base =/u01/gridexport oracle_home = $ORACLE _base/ Product/grid11gexport oracle_sid =+asm1 #节点1为 + ASM1, Node 2 is + asm2, and so on export PATH = $HOME/bin: $ORACLE _home/bin:/usr/bin:/bin:/usr/bin/x11:/usr/local/bin: $ORACLE _base/common/oracle/bin: $PATHexport ld_library_path = $ORACLE _home/lib: $ORACLE _home/oracm/lib:/usr/lib:/usr/local/libexport CLASSPATH = $ORACLE _home/jre: $ORACLE _home/jlib: $ORACLE _home/rdbms/jlib: $ORACLE _home/network/ Jlib: $ORACLE _home/oracm/lib
Oracle Environment variables: Edit the/home/oracle/.bash_profile file to add the following:
#所有节点均执行
[Email protected]:~] $cat/home/oracle/bash_profile export pathexport TMP=/tmpexport TMPDIR= $TMPexport oracle_base=/u01/oracleexport oracle_home= $ORACLE _base/product/db11gexport Oracle_sid=crrdw #此处指定oracle实例名. Oracle_sid can be specified in a single instance case, it is not recommended to specify export path =/usr/sbin in multiple instances: $PATHexport path= $ORACLE _home/bin : $PATHexport ld_library_path= $ORACLE _home/lib:/lib:/usr/libexport CLASSPATH= $ORACLE _home/jre:$ Oracle_home/jlib: $ORACLE _home/rdbms/jlib
SSH Trusted configuration (all nodes)
Configure SSH password-free login between all nodes to ensure inter-node file transfer. You need to configure SSH trust between the grid and the Oracle user separately , and the root user is not worthy. This project has only two nodes: Cheastrac01, Cheastrac02. To configure SSH Trust for a grid user, for example:
Log on to the grid user on all nodes and execute the following command:
#所有节点均执行
[Email protected]:~] $rm-rf ~/. SSH [[email protected]:~] $mkdir ~/. Ssh[[email protected]: ~/ssh[[email protected]:~]$/usr/bin/ssh-keygen-t RSA #提示输入时一路回车即可
Execute the following command on Node 1:
~/.ssh/id_rsa.pub>>~/.ssh/authorized_keys ~/.ssh/id_rsa.pub>>~/.ssh/~/.ssh/ Authorized_keys Cheastrac02:.ssh/authorized_keys
Execute the following command on Node 2:
#仅节点2执行 [[email protected]: ~/.ssh/authorized_keys
Test whether SSH Trust is configured successfully:
#所有节点均执行, the following command first executes the need to enter a password, then do not need, indicating a successful configuration [[email protected]:~] $ssh cheastrac01 date[[email protected]: ~] $ssh cheastrac02 date[[email protected]:~] $ssh cheastrac01 date[[email protected]:~] $ssh CHEASTRAC02 Date
Log in to the Oracle account and repeat the above actions.
Edit/etc/ssh/sshd_config, change #logingracetime 2m to Logingracetime 0 so that SSH authentication does not time out.
RHEL6.7 x64 Dual-node installation Oracle 11g R2 RAC