在OEL5上安装配置Oracle Gird Control 10.2.0.5

早期的Grid Control问题实在太多了，以至于把10.2.0.1的Grid Control升级到10.2.0.5几乎是不可能完成的任务；此外10.2.0.5以前的gc不支持11g作为repository database仓库数据库，不仅于此10.2.0.1版本是不支持rhel5或OEL5的，如果想安装的话rhel/oel 4是仅有理想的平台。这就这为我们制造了许多局限。如果是Fresh Installation的话似乎只安装软件(software only)，而不在安装10.2.0.1阶段配置oms，在升级到10.2.0.5后再进行oms的config会是一种比较理想的安装方法。不过我们仍需要应付一个复杂的配置过程，写这个文档的目的是帮助我们应付(包括已安装过的人，因为很容易忘记)这种窘况。 1.我们需要一个已经存在的Oracle数据库，当然它应当是纯净的(没有相关的em repository)，其版本最好是10.2.0.5或者11.2.0.2；我们假设你的数据库已经满足了一切安装grid control的前提要求，这包括设置几个初始化参数和装有dbms_shared_pool包等等。 2.其次你需要下载对应的软件，这包括了完全版本的10.2.0.1安装介质和10.2.0.5 gc patchset，并且最好有10.2.0.5版的agent(个人经验agent不太稳定，有时候需要重配，所以有介质的话会方便些):

• Linux_Grid_Control_full_102010_disk1.zip
• Linux_Grid_Control_full_102010_disk2.zip
• Linux_Grid_Control_full_102010_disk3.zip
• gc_x86_64_10205_part1of2.zip
• gc_x86_64_10205_part2of2.zip
• Linux_x86_64_Grid_Control_agent_download_10_2_0_5_0.zip

要下载这一大堆介质可能是我们安装过程中最麻烦的一件事情，特别是当你的网络状况欠佳的季节。 因为都是zip包，所以你只需要使用unzip命令将以上介质一一解压到合适目录就可以了。 3.配置OMS所在主机的内核参数和rpm包，下面给出了相关配置文件的示例值：

/etc/sysctl.conf:
kernel.shmall = 2097152
kernel.shmmax = 536870912
kernel.shmmni = 4096
kernel.sem = 250 32000 100 128
fs.file-max = 65536
# semaphores: semmsl, semmns, semopm, semmni
net.ipv4.ip_local_port_range = 1024 65000
net.core.rmem_default = 262144
net.core.rmem_max = 262144
net.core.wmem_default = 262144
net.core.wmem_max = 262144

/* 注意以上参数并不一定适合你的主机，具体如何配置请参见Metalink文档 */

同时修改/etc/security/limits.conf参数文件:
*               soft    nproc   2047
*               hard    nproc   16384
*               soft    nofile  1024
*               hard    nofile  65536

/* 星号换上你的安装用户名，如oracle或者其他dba组成员 */

安装适当的rpm包，安装database时需要的包一律也都需要，此外请特别留意安装一下几个包:
compat-libstdc++-296-2.96-138.i386
libstdc++-devel-4.1.2-48.el5.x86_64
libstdc++-devel-4.1.2-48.el5.i386
glibc-devel-2.5-49.x86_64
glibc-devel-2.5-49.i386

并建立下列符号连接:
ln -s /usr/lib/libgdbm.so.2.0.0 /usr/lib/libdb.so.2

4.上述工作完成后我们需要修改response文件以满足安装的需要，在10.2.0.1安装介质的解压目录下操作:

[root@nas media]# ls
dcommon  doc  index.htm  install  libskgxn  oms  rdbms  response  runInstaller  stage

[root@nas media]# vi response/em_using_existing_db.rsp 

/* 修改em_using_existing_db.rsp响应文件 */

包括以下参数需要从默认值修改为指定值:
UNIX_GROUP_NAME="dba"
#dba应当是有效的安装用户所在组

FROM_LOCATION="/s01/media/oms/Disk1/stage/products.xml"
#FROM LOCATION指向安装介质stage目录下的products.xml文件

BASEDIR="/s01/app/gc"
#BASEDIR指向grid control安装的基础目录

INSTALLATION_NAME="oms10g"
#安装名

s_reposHost="rh3.oracle.com"
#repository数据库的主机名或ip地址

s_reposPort="1521"
#repository数据库的监听端口

s_reposSID="EMREP"
#repository数据库的sid

s_reposDBAPwd="maclean"
#repository数据库的sys用户密码

s_mgmtTbsName="/s01/orabase/oradata/EMREP/mgmt.dbf"
#repository数据库今后的mgmt表空间的数据文件名

s_ecmTbsName=s_mgmtTbsName="/s01/orabase/oradata/EMREP/mgmt_ecm.dbf"
#repository数据库今后的ecm表空间的数据文件名

s_securePassword="maclean"
#agent将来使用的secure密码

s_securePasswordConfirm="maclean"
#确认上一步的密码

b_lockedSelected=false
#确定agent交互是否被锁

s_reposPwd="maclean"
#确定仓库数据库中模式拥有者(sysman)的密码

s_reposPwdConfirm="maclean"
#确认上一步的密码

5.以静默方式安装grid control 10.2.0.1，但不配置oms:

[maclean@nas ~]$ export  TMP=/tmp

[maclean@nas ~]$  /s01/media/install/runInstaller -noconfig -silent -ignoreSysPrereqs -responseFile \
/s01/media/response/em_using_existing_db.rsp  use_prereq_checker=false b_skipDBValidation=true -force

以上安装完成后，运行相关脚本:
[maclean@nas ~]$ su - root -c "/home/maclean/oraInventory/orainstRoot.sh"

[maclean@nas ~]$ su - root -c "/s01/app/gc/oms10g/allroot.sh"

并使用opmonctl命令关闭http等服务:
[maclean@nas ~]$ /s01/app/gc/oms10g/opmn/bin/opmnctl stopall
opmnctl: stopping opmn and all managed processes...

5.接下来我们需要安装grid control 10.2.0.5补丁集，同样的需要修改一个response响应文件:

[maclean@nas 10205]$ unzip /tmp/gc_x86_64_10205_part2of2.zip
Archive:  /tmp/gc_x86_64_10205_part2of2.zip
extracting: p3731593_10205_Linux-x86-64.zip

[maclean@nas 10205]$ unzip p3731593_10205_Linux-x86-64.zip
..............

[maclean@nas ~]$ cp /s01/10205/3731593/Disk1/response/patchset.rsp /s01/10205/3731593/Disk1/response/oms_patchset.rsp

/* 对oms_patchset.rsp修改已有的参数为指定值 */

[maclean@nas ~]$ vi /s01/10205/3731593/Disk1/response/oms_patchset.rsp

ORACLE_HOME="/s01/app/gc/oms10g"
b_softwareonly=true
s_sysPassword="maclean"
sl_pwdInfo={ "maclean" }
oracle.iappserver.st_midtier:szl_InstanceInformation={ "maclean" }

ORACLE_HOME_NAME="oms10g"
#另外增加以上条目

[maclean@nas ~]$ /s01/10205/3731593/Disk1/runInstaller -noconfig -silent \
-responseFile /s01/10205/3731593/Disk1/response/oms_patchset.rsp

/* 以上10.2.0.5补丁安装完成后，同样需要执行root.sh */

[maclean@nas ~]$ su - root -c "/s01/app/gc/oms10g/root.sh"

6.完成上述安装后可以开始配置OMS了:

[maclean@nas ~]$ export PERL5LIB=/s01/app/gc/oms10g/perl/lib/5.6.1

[maclean@nas ~]$ /s01/app/gc/oms10g/perl/bin/perl /s01/app/gc/oms10g/sysman/install/ConfigureGC.pl \
/s01/app/gc
Base Directory: /s01/app/gc

 Starting ito execute Configuration Assistants: 

Running the configuration assistants using the following command:
/s01/app/gc/oms10g/oui/bin/runConfig.sh INV_PTR_LOC=/s01/app/gc/oms10g/oraInst.loc
ORACLE_HOME=/s01/app/gc/oms10g ACTION=configure MODE=perform
COMPONENT_XML={encap_oms.1_0_0_0_0.xml}
perform - mode is starting for action: configure

以上命令的具体形式为:
<OMS ORACLE_HOME>/perl/bin/perl <OMS ORACLE_HOME>/sysman/install/ConfigureGC.pl <Parent Directory filepath> 

配置会消耗大量的时间，建议通过.../oms10g/cfgtoollogs/oui/configActions<>.log.日志文件监控配置过程:

[root@nas oui]# cd /s01/app/gc/oms10g/cfgtoollogs/oui

[root@nas oui]# tail -f configActions2011-01-23_08-57-20-AM.log
... return status = 0 (success)
Oracle JAAS [Sun Jan 23 08:57:43 CST 2011]  $ORACLE_HOME/j2ee/home/config/jazn-data.xml is synchronized successfully to dcm repository.
Please check the log file [/s01/app/gc/oms10g/cfgtoollogs/jaznca.log] for details.

The plug-in Java Security Configuration Assistant has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in Web Cache Configuration Assistant is running

2
Start traversing...
got process-manager node
got ias-instance node
attrValue=IASPT

attrValue=DSA

attrValue=HTTP_Server

attrValue=LogLoader

attrValue=dcm-daemon

attrValue=OC4J

attrValue=WebCache

Entity found.

got ias-instance node
Current status is : enabled
Changing the value of port to enabled
 Modified ...
Before After WaitForComplete
After WaitForComplete
Completed smiSetStatus
Checking status ... enableconfiguration
In ... getWebcachePort
WebCache Default Port :7777
In ... updateApacheConf
Apache Port Value : 7777
Apache Before WaitForComplete
Apache After WaitForComplete
Checking for Apache updation status
Apache httpd.conf updated
smiTearDown
Will be checking the status ...
Webcache Configuration finished successfully

The plug-in Web Cache Configuration Assistant has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in Oracle Application Server Instance Configuration Assistant is running

The plug-in Oracle Application Server Instance Configuration Assistant has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in OC4J Instance Configuration Assistant is running

Reading ini file - '/s01/app/gc/oms10g/j2ee/deploy.ini'
Adding web-app 'IsWebCacheWorkingWeb.war' for app 'IsWebCacheWorking'.
Adding web-app 'wsrp-samples.war' for app 'portletapp'.
Initializing DCM...done.
OC4J instance 'home' already exists.
Starting OC4J instance 'home'...done.
Deploying application 'portletapp' to OC4J instance 'home'.
Notification ==> Application Deployer for portletapp STARTS [ 2011-01-23T08:58:00.972CST ]
Notification ==> Undeploy previous deployment
Notification ==> Removing files for app file:/s01/app/gc/oms10g/j2ee/home/applications/portletapp.ear
Notification ==> Copy the archive to /s01/app/gc/oms10g/j2ee/home/applications/portletapp.ear
Notification ==> Unpack portletapp.ear begins...
Notification ==> Unpack portletapp.ear ends...
Notification ==> Initialize portletapp.ear begins...
Notification ==> Initialize portletapp.ear ends...
Notification ==> Initialize wsrp-samples begins...
Notification ==> Initialize wsrp-samples ends...
Notification ==> deleting:  /s01/app/gc/oms10g/j2ee/home/applications/portletapp.ear
Notification ==> deleting:  /s01/app/gc/oms10g/j2ee/home/applications/portletapp/wsrp-samples.war
Notification ==> Application Deployer for portletapp COMPLETES [ 2011-01-23T08:58:01.319CST ] 

Deploying application 'IsWebCacheWorking' to OC4J instance 'home'.
Notification ==> Application Deployer for IsWebCacheWorking STARTS [ 2011-01-23T08:58:01.328CST ]
Notification ==> Undeploy previous deployment
Notification ==> Removing files for app file:/s01/app/gc/oms10g/j2ee/home/applications/IsWebCacheWorking.ear
Notification ==> Copy the archive to /s01/app/gc/oms10g/j2ee/home/applications/IsWebCacheWorking.ear
Notification ==> Unpack IsWebCacheWorking.ear begins...
Notification ==> Unpack IsWebCacheWorking.ear ends...
Notification ==> Initialize IsWebCacheWorking.ear begins...
Notification ==> Initialize IsWebCacheWorking.ear ends...
Notification ==> Initialize IsWebCacheWorkingWeb begins...
Notification ==> Initialize IsWebCacheWorkingWeb ends...
Notification ==> deleting:  /s01/app/gc/oms10g/j2ee/home/applications/IsWebCacheWorking.ear
Notification ==> deleting:  /s01/app/gc/oms10g/j2ee/home/applications/IsWebCacheWorking/IsWebCacheWorkingWeb.war
Notification ==> Application Deployer for IsWebCacheWorking COMPLETES [ 2011-01-23T08:58:01.362CST ] 

Calling updateConfig to notify DCM of new deployments...done.
Adding dependent libraries for application 'portletapp'...done.
Adding OC4J mount points for application 'portletapp'...done.
Adding OC4J mount points for application 'IsWebCacheWorking'...done.
Calling SMI to save changes.
SMISession.saveChanges succeeded.
Binding web app 'wsrp-samples' to default-web-site for application 'portletapp' in OC4J instance 'home'
Web app 'wsrp-samples' bound successfully.
Binding web app 'IsWebCacheWorkingWeb' to default-web-site for application 'IsWebCacheWorking' in OC4J instance 'home'
Web app 'IsWebCacheWorkingWeb' bound successfully.
Calling updateConfig to notify DCM of new web-bindings...done.
Adding application 'portletapp' to the DCM repository...done.
Application 'portletapp' deployed successfully.
Adding application 'IsWebCacheWorking' to the DCM repository...done.
Application 'IsWebCacheWorking' deployed successfully.
Stopping OC4J instance 'home'...done.
Calling SMI to retry init of failed plugins...done.
Terminating DCM...done.
Copying /s01/app/gc/oms10g/j2ee/deploy.ini to /s01/app/gc/oms10g/j2ee/deploy.ini.1295744298019.bak.
Writing any undeployed entries back to /s01/app/gc/oms10g/j2ee/deploy.ini.

Oc4jDeploy tool completed successfully!

The plug-in OC4J Instance Configuration Assistant has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in Register DCM Plug-Ins With EM is running

Operation successful.

The plug-in Register DCM Plug-Ins With EM has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in DCM Repository Backup Assistant is running

backup created: InstalledImage_EnterpriseManager0.nas

The plug-in DCM Repository Backup Assistant has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in EM Technology Stack Upgrade is running

The plug-in EM Technology Stack Upgrade has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in OMS Configuration is running

Operation Stopping OPMN Processes is in progress.
Operation EM Deploying is in progress.
Operation Creating OMS Respository is in progress.
Operation Configuring OMS is in progress.
OMS is being Secured and Lock is set to false.
Precompiling JSPs.
Performing installation of CLI services for client.
Operation Restarting OPMN Processes is in progress.

The plug-in OMS Configuration has successfully been performed
------------------------------------------------------
------------------------------------------------------
The plug-in Agent Configuration Assistant is running

Performing free port detection on host=nas
Securing the agent
Performing targets discovery and agent configuration

The plug-in Agent Configuration Assistant has failed its perform method
------------------------------------------------------
The action configuration has failed its perform method
###################################################

7.如果以上oms configuration顺利完成那么Grid Control的网页界面已经可以登录了，但我们在本地服务器上的agent仍是10.2.0.1版本的(所以Agent Configuration Assistant失败了)，这里我推荐将原agent目录删除后通过10.2.0.5的agent介质(Linux_x86_64_Grid_Control_agent_download_10_2_0_5_0.zip)重新安装并部署，这样可以很大程度上避免出现问题。