原grid

ORACLE_BASE=/u01/app/oracle

ORACLE_HOME=/u01/app/grid

原oracle

ORACLE_BASE=/u01/app/oracle

ORACLE_ HOME =/u01/app/oracle/product/11.2.0/db_1

新grid

ORACLE_BASE=/u01/app/oracle

ORACLE_HOME=/u01/app/110204/grid

新oracle

ORACLE_BASE=/u01/app/oracle

ORACLE_ HOME =/u01/app/oracle/product/11.2.0.4/db_1

1．上传新版本oracle和grid安装文件至第一个节点，并解压出来

2．所有节点上新建目录/u01/app/110204/grid和/u01/app/oracle/product/11.2.0.4/db_1并分别授予grid和oracle的权限

3． Grid升级前检查，进入grid安装文件的目录，执行如下语句（-src_crshome后面对应的是原来grid的$ORACLE_HOME, -dest_crshome后面对应的是新版本grid的$ORACLE_HOME，个人实验的情况是两个问题都忽略了）

./runcluvfy.sh stage -pre crsinst -upgrade -n testdb01,testdb02 -rolling -src_crshome /u01/app/grid -dest_crshome /u01/app/110204/grid -dest_version 11.2.0.4.0 -fixup -fixupdir /tmp -verbose

一般都有如下两个问题(第一个解决办法：安装9706490补丁包;第二个解决办法：忽略)

Checking for Oracle patch "9413827 or 9706490" in home "/u01/app/grid".

Node Name Applied Required Comment

-------- ------ ------ ---

testdb02 missing 9413827 or 9706490 failed

testdb01 missing 9413827 or 9706490 failed

Result: Check for Oracle patch "9413827 or 9706490" in home "/u01/app/grid" failed

There are no oracle patches required for home "/u01/app/110204/grid".

Checking if search entry in file "/etc/resolv.conf" is consistent across the nodes...

search entry in file "/etc/resolv.conf" is consistent across nodes

Checking DNS response time for an unreachable node

Node Name Status

------- ----

testdb02 failed

testdb01 failed

PRVF-5636 : The DNS response time for an unreachable node exceeded "15000" ms on following nodes: testdb02,testdb01

File "/etc/resolv.conf" is not consistent across nodes

4．关闭所有节点的CRS（个人实验过程中因为没有打9706490补丁，所以采用了关闭CRS来避免后期执行rootupgrade.sh导致Failed to stop the lower version of Grid Infrastructure错引发的grid升级不成功的问题）

Root#/u01/app/grid/bin/crsctl stop crs

5．安装grid

6．启动所有节点的CRS(当然安装完grid后，过一段时间发现会自动启动grid即crsctl start crs，但是安装过程中还是手工启动的好，且启动后的grid的版本有更新了如下)

Root#/u01/app/grid/bin/crsctl start crs

crsctl query crs activeversion -- 11.2.0.4.0

/u01/app/110204/grid/bin/crsctl query crs activeversion -- 11.2.0.4.0

crsctl query crs releaseversion -- 11.2.0.1.0

/u01/app/110204/grid/bin/crsctl query crs releaseversion -- 11.2.0.4.0

crsctl query crs softwareversion -- 11.2.0.4.0

/u01/app/110204/grid/bin/crsctl query crs softwareversion -- 11.2.0.4.0

srvctl config asm显示ASM home: /u01/app/110204/grid

/u01/app/110204/grid/bin/srvctl config asm显示ASM home: /u01/app/110204/grid

但是grid用户下sqlplus / as sysasm显示实例时idle状态，asmcmd进入也是idle状态，执行srvctl status asm却显示ASM is running。查看了$ORACLE_BASE/diag/asm目录没有发现另外的ASM实例名称，也就是ASM实例的名称还是低版本的，那为何grid用户下sqlplus / as sysasm显示idle，但是oracle用户下sqlplus / as sysdba可以查询，奇怪了，理论上ASM没有启动，DB是启动不了的。

7．安装oracle，选择update an existing databae，执行完$ORACLE_HOME/root.sh后会自动运行DBUA，这些步骤完成后不用再手动执行什么catupgrd.sql、utlrp.sql了，因为DBUA的过程中会自动performing post upgrade之类