vmware上安装 ole 4+ora10g rac的问题.docx

1、vmware上安装 ole 4+ora10g rac的问题因为vmware workstation本身不支持双机，要实现clusterware的ocr、voting盘，需要以下文件系统：ocfs/ocfs2raw因此安装clusterware之前需要安装支持双机的文件系统ocfs2On Linux platforms, ASM can use raw devices or devices managed via the ASMLib interface. Oracle recommends ASMLib over raw devices for ease-of-use and performa

2、nce reasons. ASMLib 2.0 is available forfree downloadfrom OTN. This section walks through the process of configuring a simple ASM instance by using ASMLib 2.0 and building a database that uses ASM for disk storage.1、创建虚拟机a2、虚拟机a添加裸设备的硬盘：依次添加ocr.vmdk、voting.vmdk、asm1.vmdk、asm2.vmdk 并修改为SCSI 1:03、修改oe

3、l4a.vmx，disk.locking = FALSEdiskLib.dataCacheMaxSize = 0 diskLib.dataCacheMaxReadAheadSize = 0 diskLib.dataCacheMinReadAheadSize = 0 diskLib.dataCachePageSize = 4096 diskLib.maxUnsyncedWrites = 0scsi1.present = TRUEscsi1.virtualDev = lsilogicscsi1.sharedBus = VIRTUALscsi1:0.present = TRUEscsi1:0.fil

4、eName = D:VMWAREsharedisk10gocr.vmdkscsi1:0.mode = independent-persistentscsi1:0.deviceType = diskscsi1:1.present = TRUEscsi1:1.fileName = D:VMWAREsharedisk10gvoting.vmdkscsi1:1.mode = independent-persistentscsi1:1.deviceType = diskscsi1:2.present = TRUEscsi1:2.fileName = D:VMWAREsharedisk10gasm1.vm

5、dkscsi1:2.mode = independent-persistentscsi1:2.deviceType = diskscsi1:3.present = TRUEscsi1:3.fileName = D:VMWAREsharedisk10gasm2.vmdkscsi1:3.mode = independent-persistentscsi1:3.deviceType = disk4、增加一个网卡：5、启动虚拟机，开始安装oel4操作系统选择分区： ；接下来的告警，选Y将/dev/sda分区如下：Mount point size type 备注/ 7GB ext3 选中”force t

6、o be a primary partition” 2GB swap /u01 剩余 ext3 选择禁用防火墙和SElinux选择“Customize software packages to be installed”X Window SystemGnome Desktop EnvironmentEditorsGraphical InternetServer Configuration ToolsLegacy Network (click Details and select “rsh_server” and “telnet-server”)Development ToolsLegacy S

7、oftware DevelopmentAdministration ToolsSystem Tools (select all packages that starts with “ocfs2” and “oracleasm”, select “systat” as well)安装完毕后，安装vmwaretools，解压缩安装包后，执行.vmware-config-tools.plvmware-toolbox-cmd timesync enable;Edit /boot/grub/grub.conf file and add “clock=pit nosmp noapic nolapic” t

8、o the line that reads kernel /bootThe “clock=pit” prevents the clock for running to quickly and “nosmp noapic nolapic” prevents the clock from running too slowly. 然后重启操作系统，使上述配置生效。6、格式化共享盘disk /dev/sdbType n (to create new partition)Type p to create a primary partitionType 1 to provide the partition

9、 numberDouble click on Enter and type “w” to save the changesPerform above steps for all hard disk:fdisk /dev/sdcfdisk /dev/sddfdisk /dev/sde7、重复上述步骤1-5，安装虚拟机b其中共享磁盘不需要在界面添加，直接编辑oel4b.vmx，增加以下内容disk.locking = FALSEdiskLib.dataCacheMaxSize = 0 diskLib.dataCacheMaxReadAheadSize = 0 diskLib.dataCacheMi

10、nReadAheadSize = 0 diskLib.dataCachePageSize = 4096 diskLib.maxUnsyncedWrites = 0scsi1.present = TRUEscsi1.virtualDev = lsilogicscsi1.sharedBus = VIRTUALscsi1:0.present = TRUEscsi1:0.fileName = D:VMWAREsharedisk10gocr.vmdkscsi1:0.mode = independent-persistentscsi1:0.deviceType = diskscsi1:1.present

11、= TRUEscsi1:1.fileName = D:VMWAREsharedisk10gvoting.vmdkscsi1:1.mode = independent-persistentscsi1:1.deviceType = diskscsi1:2.present = TRUEscsi1:2.fileName = D:VMWAREsharedisk10gasm1.vmdkscsi1:2.mode = independent-persistentscsi1:2.deviceType = diskscsi1:3.present = TRUEscsi1:3.fileName = D:VMWAREs

12、haredisk10gasm2.vmdkscsi1:3.mode = independent-persistentscsi1:3.deviceType = disk7、添加用户、设置环境变量见orainst.txt8、添加裸设备map裸设备, 修改 /etc/sysconfig/rawdevices文件，添加:/dev/raw/raw1 /dev/sdb1/dev/raw/raw2 /dev/sdc1执行/sbin/service rawdevices restart 使其生效修改裸设备的权限:chown oracle:dba /dev/raw/raw*chmod 660 /dev/raw/r

13、aw*ls -lat /dev/raw/raw*root用户，修改/etc/udev/permissions.d/50-udev.permissions 文件，添加：# raw devicesram*:root:disk:0660#raw/*:root:disk:0660raw/*:oracle:dba:06609、创建双机ssh互信任双机节点上执行：su oraclessh-keygen t rsassh-keygen t dsa将生成文件拷贝到 /.ssh/authorized_keys，在节点1执行：cat /.ssh/id_rsa.pub /.ssh/authorized_keysca

14、t /.ssh/id_dsa.pub /.ssh/authorized_keysssh rac2 cat /.ssh/id_rsa.pub /.ssh/authorized_keysssh rac2 cat /.ssh/id_dsa.pub /.ssh/authorized_keysscp /.ssh/authorized_keys rac2:/.ssh/authorized_keys执行以下两次，进行测试（第一次输入yes，后续无提示直接输出日期结果）ssh rac1 datessh rac2 datessh rac1-priv datessh rac2-priv date10、配置orac

15、leasm 节点1上oracleasm configoralceasm createdisk VOL1 /dev/sdd1oralceasm createdisk VOL2 /dev/sde1节点2上：Oracleasm scandisks11、安装oracle clusterwareroot用户下：xhost +切换到oracle用户开始安装./runInstaller按照提示在root用户在两个节点分别执行orainstRoot.sh（逐个节点执行）在root用户在两个节点分别执行root.sh（逐个节点执行）执行完毕后，注意先不要返回clusterware安装界面点finish，用roo

16、t用户执行vipca进行配置（因为我们配置的eth0的ip 192.168.2.*为私网IP ，安装程序未自动识别进行VIPCA时导致失败）cd /u01/app/oracle/product/10.2.0/crs_1/bin./vipca输入节点1的 “IP Alias Name”后，其他自动关联返回clusterware安装界面，finish12、安装oracle 数据库软件选择只安装数据库软件按照提示在节点1、节点2上分别执行root.sh13、进行netca配置在节点1 oracle用户下，执行netca配置listener配置后终端窗口输出：执行crs_stat t查看，两个节点上的

17、lsnr服务启动14、创建ASM实例及数据库dbca不选择OEM选择IFILE选择OK“Change Disk Discovery Paht”为”/dev/oracleasm/disks/*”创建DG弹出报错，先不管它类似的创建FG，然后在关闭节点2、节点1 ，重新启动节点1、节点2，并root用户先后执行/etc/init.d/init.crs stop/etc/init.d/init.crs start然后在dbca界面cancle，重新运行dbca，并选择配置ASM可以看到两个磁盘组在2个节点上都mount上来了继续运行dbca，创建数据库选择不要创建OEM，可以后续手工创建选择磁盘组D

18、G“Specify Flash Recovery Area”，选择磁盘组FG选择字符集：点击Finish，完成数据库创建：执行crs_stat -t，运行结果输出如下，所以服务正常15、安装clusterware 10.2.0.5补丁 （clusterware补丁的安装软件和oralce的是一个安装包）首先停掉以下服务srvctl stop database -d devdbsrvctl stop asm -n rac1srvctl stop asm -n rac2srvctl stop listener -n rac1srvctl stop listener -n rac2之后，执行crs_

19、stat t查看各个服务状态如下：OK，开始安装补丁，选择HOME为”OraCrs10g_home”，其他默认即可，开始打补丁根据提示，先后在节点1、节点2上root用户执行以下步骤： 1、$ORA_CRS_HOME/bin/crsctl stop crs2、$ORA_CRS_HOME/install/root102.sh此时,检查crs状态，发现ora.rac2.vip的状态是OFFLINE，重启动几次 crs，状态仍然如此。检查$ORA_CRS_LOG/log/racg/ora.rac1.vip.log，发现以下报错：2013-07-17 11:22:46.797: RACG4143876

20、864 105774143876864ora.rac1.vip: ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1)ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1)2013-07-17 11:22:53.992: RACG4124580608 107644124580608ora.rac1.vip: clsrcstartorp: Error with malloc2013-07-17 11:23:38.677: RACG4143876864 110174143876864o

21、ra.rac1.vip: ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1)ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1)Interface eth0 checked failed (host=rac1)ping to 192.168.2.1 via eth1 failed, rc = 1 (host=rac1)ping to 192.168.2.1 via eth1 failed, rc = 1 (2013-07-17 11:23:38.678: RACG4143876

22、864 110174143876864ora.rac1.vip: host=rac1)虚拟机设置的默认网关为192.168.2.1，是笔记本电脑的vnet8虚拟网卡的网络地址，因为本机为win8，默认关闭了ICMP V4，所以，在虚拟机上无法ping通该网关。开启方法，在windows防火墙中开启：然后在节点1、节点2分别重启动CRS服务，各个服务状态恢复正常：16、安装oracle database 10.2.0.5补丁 （oralce database补丁的安装软件和clusterware的是一个安装包）首先停掉以下服务srvctl stop database -d devdbsrvctl

23、 stop asm -n rac1srvctl stop asm -n rac2srvctl stop listener -n rac1srvctl stop listener -n rac2运行crs_stat t 检查服务状态如下，就可以开始打补丁了运行./runInstaller开始打补丁,注意在HOME选择列表中选择ORACLE DATABASE的HOME名”OraDb10g_home1”后面继续即可，直到install完成，需要手工执行脚本：依次在节点1、几点2上以root用户执行脚本创建数据库后，升级10.2.0.1- 10.2.0.5报错如下：ORA-00704: bootstr

24、ap process failureORA-39700: database must be opened with UPGRADE option参考metalink案例【ID 435337.1】，因为在升级DATABASE时没有按照README的Postinstallation tasks说明进行操作，两种方法：1、使用Oracle Database Upgrade Assistant(dbua)，以oracle用户执行 dbua ，然后按照提示进行升级操作即可。2、手工执行升级，参见补丁包的README.HTML，比较比较繁琐，建议还是dbua吧检查两个节点的listener服务运行正常，执行dbua，选择升级”Database”，获取数据库信息这一步比较慢，耐心等候选择在升级的最后步骤编译失效对象升级过程超慢，耐心等待，可能是虚拟机上性能较差最后显示升级结果如下：