RAC安装配置注意文档(Linux版)

工作环境：

Microsoft Windows XP Professional Service Pack 3 Vmware版本：Vmware GSX Server 3.2.1 build-19281 Linux 版本：Oracle Enterprise-R5-U2-Server crs:10201_clusterware_linux32.zip

oracle_home:10201_database_linux32.zip

关于PUBLIC NODE NAME

和机器名相同，机器名应较短（原因详见Failed to upgrade Oracle Cluster Registry configuration报错）[修改机器名：/etc/sysconfig/network]

网络组件配置

多个虚拟机Private NIC引用同一个Vmnet(n)虚拟网卡(启动DHCP服务) 然后使用固定IP不同的服务器绑定不同的地址。

使用Vmware创建共享磁盘 创建磁盘：

vmware-vdiskmanager -c -s 10Gb -a E:\\vm_linuxs\\linuxserver1\\shareddisk.vmdk

lsilogic

-t

0

设置共享：

多个虚拟机文件的vmx都需要修改同样配置： scsi1.present = \"TRUE\" scsi1:0.present = \"TRUE\"

scsi1:0.fileName = \"E:\\vm_linuxs\\linuxserver1\\shareddisk.vmdk\" scsi1:0.mode = \"independent-persistent\" scsi1.virtualDev = \"lsilogic\" disk.locking=\"FALSE\"

scsi1.SharedBus=\"Virtual\"

相关设置：

OCR,VDISK裸设备配置 磁盘分区：

fdisk /dev/sdb

/dev/sdb5 (100M) /dev/sdb6 (50M)

添加映射文件：

raw /dev/raw/raw1 /dev/sdb5 (OCR) raw /dev/raw/raw2 /dev/sdb6 (Vdisk)

随系统自动映射：

系统重启后raw后的映射文件会消失，所以需要在启动时完成映射工作(所有NODE)： /etc/rc.local

raw /dev/raw/raw1 /dev/sdb5 raw /dev/raw/raw2 /dev/sdb6

chown oracle:oinstall /dev/raw/*

为用户等效性配置 SSH 每台服务器以Oracle身份执行：

mkdir ~/.ssh

chmod 755 ~/.ssh

/usr/bin/ssh-keygen -t rsa /usr/bin/ssh-keygen -t dsa

主服务器以Oracle身份执行：

cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

ssh oracle@ds2 cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys ssh oracle@ds2 cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys chmod 644 ~/.ssh/authorized_keys

次服务器以Oracle身份执行：

cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

ssh oracle@ds1 cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys ssh oracle@ds1 cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys chmod 644 ~/.ssh/authorized_keys

测试连通性

所有的机器都要测试，临时性，下次需要重新执行 exec /usr/bin/ssh-agent $SHELL /usr/bin/ssh-add ssh ds2 date

确保Oracle用户之间的连通有效性

在执行集群安装(./runInstaller,dbca,vipca等)前要执行： ssh date

ssh date ssh date ssh date

错误收集

执行root.sh报错: Failed to upgrade Oracle Cluster Registry configuration

there is an orable bug causally for the described problem: bug4632899: CSS not starting up if hostname has capital letters - OS Solaris After changing hostnames to smaller case we retried CRS installation successfully.

执行root.sh的Oracle Cluster Verification Utility Failed

(是由于root.sh调用vipca执行失败，详情见下)

执行vipca时报错:libpthread.so.0: cannot open shared object file :

SRVCTL and VIPCA Utilities Set the LD_ASSUME_KERNEL Parameter The SRVCTL and VIPCA utilities shipped with Oracle Database 10g release 2 and Oracle Clusterware software set the environmental variable LD_ASSUME_KERNEL. On SUSE Linux Enterprise Server 10, because the older Linux threads API has been removed from GLIBC, setting this parameter causes the SRVCTL and VIPCA utilities to exit with the following error: /opt/oracle/crs/jdk/jre/bin/java:

error while loading shared libraries:

libpthread.so.0: cannot open shared object file: No such file or directory

Workaround: Comment out the lines that set the LD_ASSUME_KERNEL variable from the VIPCA and SRVCTL utilities. For the VIPCA utility alter the $CRS_HOME/bin/vipca script commenting out lines 119 through 123 as follows: arch='uname -m' # if [ \"$arch\" = \"i686\" -o \"$arch\" = \"ia64\" -o \"$arch\" = \"x86_64\" ] # then

# LD_ASSUME_KERNEL=2.4.19 # export LD_ASSUME_KERNEL # fi

With the lines commented out, root.sh should be able to call VIPCA successfully. Ensure that you do not to comment out line 118 which sets the arch variable as that is needed by the script.

For the SRVCTL utility alter the $CRS_HOME/bin/srvctl and the $ORACLE_HOME/bin/srvctl scripts commenting out lines 173 and 174 as follows:

#Remove this workaround when the bug 3937317 is fixed #LD_ASSUME_KERNEL=2.4.19 #export LD_ASSUME_KERNEL

执行vipca时报错:Running vipca(silent) for configuring nodeapps

Error 0(Native: listNetInterfaces:[3])

[Error 0(Native: listNetInterfaces:[3])]

安装CRS过程中，请点击提示执行root.sh的窗口完成，进入最后的配置验证界面，然

后再单独运行vipca(root图形化)。

执行root.sh时报错：Oracle CRS stack is already configured and will be running under init(1M)

导致失败：

解决办法：

在每个节点上删除文件：/etc/oracle/scls_scr/rac1/oracle/cssfatal

其他

root.sh在次服务器上执行成功