五、总结 － 及其他说明/TIPS/文中提到的有用脚本
：
从上面的配置及测试我们可以看到，只要掌握了应用程序的启动、关闭、检查的方法，我们就完全可以
利用Oracle CRS来搭建一般应用程序的高可用集群，从而保证了业务的连续性。
1.简单错误处理：如果在实施或测试过程中命令失败，或者有错误需要回退状态，
或者删除资源及相关信息，可以使用下面的几个命令：
crs_stop -f db
crs_unregister db
crs_profile -delete db
错误二：我在上面启动虚拟IP资源的时候报错：
root@rac01 # crs_start –all
….
Attempting to start `havip` on member `rac01`
`havip` on member `rac01` has experienced an unrecoverable failure.
Human intervention required to resume its availability.
…..
查看crs日志：
这时候发现在输入网卡时候不小心输入了错误的网卡名称，公共网卡名称应该是e1000g0，
结果使用的bge0，使用下面命令修改
root@rac01 # crs_profile -update havip -o oi=e1000g0
再次启动资源成功。
2.文中的sysdba命令是我习惯用的alias，需要在oracle的.bash_profile或者.profile中添加：
alias sysdba="sqlplus '/as sysdba'"
文中的命令crs_start/crs_stop/crs_profile/crs_register等命令位于$CRS_HOME/bin/目录下，
为了使用方面，一般需要手工将该PATH添加到root和oracle的PATH环境变量中去。
3.磁盘组的启动、关闭、监测脚本db.sh
4.文件系统的启动、关闭、监测脚本fs.sh
#!/bin/sh
# *****************************************************************
# shahand 2008-1-3
SCRIPT=$0
ACTION=$1 # Action (start, stop or check)
DG_NAME=oradata12
VOL_NAME=oradata
case $1 in
'start')
mount -F vxfs /dev/vx/dsk/$DG_NAME/$VOL_NAME /oradata
echo "Resource STARTED"
;;
'stop')
umount /oradata
echo "Resource STOPPED"
;;
'check')
if [[ `mount |grep $VOL_NAME |wc -c` -eq 0 ]]
then
exit 1
else
exit 0
fi
echo "Resource CHECKED"
;;
*)
echo "usage: $0 {start stop check}"
;;
esac
exit 0
5. 数据库的启动、关闭、监测脚本db.sh
#!/bin/sh
# *****************************************************************
# ## shahand 2008-1-3
# *****************************************************************
# /oracle/crs/crs/public/db.sh
SCRIPT=$0
ACTION=$1 # Action (start, stop or check)
CRS_HOME=/oracle/crs/
case $1 in
'start')
su - oracle -c "export ORACLE_SID=orcl;
sqlplus /nolog <<EOF
connect /as sysdba;
startup;
exit;
EOF"
exit $?
;;
'stop')
su - oracle -c "export ORACLE_SID=orcl;
sqlplus /nolog <<EOF1
connect /as sysdba;
alter system checkpoint;
shutdown immediate;
exit;
EOF1"
exit $?
;;
'check')
if [ ! -f $CRS_HOME/crs/public/db.sql ]
then
echo "connect /as sysdba;" > $CRS_HOME/crs/public/db.sql;
echo "spool /tmp/db_check.out;" >>$CRS_HOME/crs/public/db.sql;
echo "select status from v/$instance ;" >> $CRS_HOME/crs/public/db.sql;
echo "exit;" >> $CRS_HOME/crs/public/db.sql;
fi
su - oracle -c "export ORACLE_SID=orcl;sqlplus /nolog @$CRS_HOME/crs/public/db.sql";
if [ `cat /tmp/db_check.out|grep OPEN|wc -c` -eq 0 ]
then
exit 1
else
exit 0
fi
;;
*)
echo "usage: $0 {start stop check}"
;;
esac
exit 0
5. 监听的启动、关闭、监测脚本lsnr.sh
7.Oracle集群软件的安装响应文件crs.rsp ，其中在####之前需要修改
从上面的配置及测试我们可以看到，只要掌握了应用程序的启动、关闭、检查的方法，我们就完全可以
利用Oracle CRS来搭建一般应用程序的高可用集群，从而保证了业务的连续性。
1.简单错误处理：如果在实施或测试过程中命令失败，或者有错误需要回退状态，
或者删除资源及相关信息，可以使用下面的几个命令：
crs_stop -f db
crs_unregister db
crs_profile -delete db
错误二：我在上面启动虚拟IP资源的时候报错：
root@rac01 # crs_start –all
….
Attempting to start `havip` on member `rac01`
`havip` on member `rac01` has experienced an unrecoverable failure.
Human intervention required to resume its availability.
…..
查看crs日志：
<!--
Code highlighting produced by Actipro CodeHighlighter (freeware)
http://www.CodeHighlighter.com/
-->
2008
-
01
-
10
17
:
30
:
22.526
:
[
CRSRES
][
1580
]
startRunnable: setting CLI
values
2008
-
01
-
10
17
:
30
:
22.527
:
[
CRSRES
][
1580
]
Attempting
to
start `dg`
on
member `rac01`
2008
-
01
-
10
17
:
30
:
22.589
:
[
CRSRES
][
1581
]
startRunnable: setting CLI
values
2008
-
01
-
10
17
:
30
:
22.629
:
[
CRSRES
][
1581
]
Attempting
to
start `havip`
on
member `rac01`
2008
-
01
-
10
17
:
30
:
22.688
:
[
CRSAPP
][
1581
]
StartResource error
for
havip error code
=
1
2008
-
01
-
10
17
:
30
:
22.749
:
[
CRSAPP
][
1581
]
StopResource error
for
havip error code
=
1
2008
-
01
-
10
17
:
30
:
22.757
:
[
CRSRES
][
1581
]
X_OP_StopResourceFailed : Stop Resource failed (
File
: rti.cpp, line:
1796
2008
-
01
-
10
17
:
30
:
22.758
:
[
CRSRES
][
1581
][
ALERT
]
`havip`
on
member `rac01` has experienced an ...
2008
-
01
-
10
17
:
30
:
22.758
:
[
CRSRES
][
1581
]
Human intervention required
to
resume its availability.
2008
-
01
-
10
17
:
30
:
23.211
:
[
CRSRES
][
1580
]
Start
of
`dg`
on
member `rac01` succeeded.
结果使用的bge0，使用下面命令修改
root@rac01 # crs_profile -update havip -o oi=e1000g0
再次启动资源成功。
2.文中的sysdba命令是我习惯用的alias，需要在oracle的.bash_profile或者.profile中添加：
alias sysdba="sqlplus '/as sysdba'"
文中的命令crs_start/crs_stop/crs_profile/crs_register等命令位于$CRS_HOME/bin/目录下，
为了使用方面，一般需要手工将该PATH添加到root和oracle的PATH环境变量中去。
3.磁盘组的启动、关闭、监测脚本db.sh
<!--
Code highlighting produced by Actipro CodeHighlighter (freeware)
http://www.CodeHighlighter.com/
-->
#!
/
bin
/
sh #
*****************************************************************
# shahand
2008
-
1
-
3
SCRIPT
=
$
0
ACTION
=
$
1
# Action (start, stop
or
check
) DG_NAME
=
oradata12 VOL_NAME
=
oradata
case
$
1
in
'
start
'
)
/
usr
/
sbin
/
vxdg
-
tfC import $DG_NAME;
/
usr
/
sbin
/
vxvol
-
g $DG_NAME startall echo "Resource STARTED" ;;
'
stop
'
)
/
usr
/
sbin
/
vxvol
-
g $DG_NAME stopall;
/
usr
/
sbin
/
vxdg deport $DG_NAME echo "Resource STOPPED" ;;
'
check
'
)
if
[
[ `/usr/sbin/vxprint -v -g $DG_NAME |grep $VOL_NAME|wc -c` -eq 0
]
]
then
exit
1
else
exit
0
fi echo "Resource CHECKED" ;;
*
) echo "usage: $
0
{start stop
check
}" ;; esac
exit
0
4.文件系统的启动、关闭、监测脚本fs.sh
#!/bin/sh
# *****************************************************************
# shahand 2008-1-3
SCRIPT=$0
ACTION=$1 # Action (start, stop or check)
DG_NAME=oradata12
VOL_NAME=oradata
case $1 in
'start')
mount -F vxfs /dev/vx/dsk/$DG_NAME/$VOL_NAME /oradata
echo "Resource STARTED"
;;
'stop')
umount /oradata
echo "Resource STOPPED"
;;
'check')
if [[ `mount |grep $VOL_NAME |wc -c` -eq 0 ]]
then
exit 1
else
exit 0
fi
echo "Resource CHECKED"
;;
*)
echo "usage: $0 {start stop check}"
;;
esac
exit 0
5. 数据库的启动、关闭、监测脚本db.sh
#!/bin/sh
# *****************************************************************
# ## shahand 2008-1-3
# *****************************************************************
# /oracle/crs/crs/public/db.sh
SCRIPT=$0
ACTION=$1 # Action (start, stop or check)
CRS_HOME=/oracle/crs/
case $1 in
'start')
su - oracle -c "export ORACLE_SID=orcl;
sqlplus /nolog <<EOF
connect /as sysdba;
startup;
exit;
EOF"
exit $?
;;
'stop')
su - oracle -c "export ORACLE_SID=orcl;
sqlplus /nolog <<EOF1
connect /as sysdba;
alter system checkpoint;
shutdown immediate;
exit;
EOF1"
exit $?
;;
'check')
if [ ! -f $CRS_HOME/crs/public/db.sql ]
then
echo "connect /as sysdba;" > $CRS_HOME/crs/public/db.sql;
echo "spool /tmp/db_check.out;" >>$CRS_HOME/crs/public/db.sql;
echo "select status from v/$instance ;" >> $CRS_HOME/crs/public/db.sql;
echo "exit;" >> $CRS_HOME/crs/public/db.sql;
fi
su - oracle -c "export ORACLE_SID=orcl;sqlplus /nolog @$CRS_HOME/crs/public/db.sql";
if [ `cat /tmp/db_check.out|grep OPEN|wc -c` -eq 0 ]
then
exit 1
else
exit 0
fi
;;
*)
echo "usage: $0 {start stop check}"
;;
esac
exit 0
5. 监听的启动、关闭、监测脚本lsnr.sh
<!--
Code highlighting produced by Actipro CodeHighlighter (freeware)
http://www.CodeHighlighter.com/
-->
#!
/
bin
/
sh #
*****************************************************************
# ## shahand
2008
-
1
-
3
#
*****************************************************************
#
/
oracle
/
crs
/
crs
/
public
/
lsnr.sh SCRIPT
=
$
0
ACTION
=
$
1
# Action (start, stop
or
check
)
case
$
1
in
'
start
'
) su
-
oracle
-
c "lsnrctl start"
exit
$? ;;
'
stop
'
) su
-
oracle
-
c "lsnrctl stop"
exit
$? ;;
'
check
'
)
if
[
`ps -ef|grep tnslsnr |wc -c` -eq 0
]
then
echo "bad";
exit
1
else
echo "good";
exit
0
fi ;;
*
) echo "usage: $
0
{start stop
check
}" ;; esac
exit
0
<!--
Code highlighting produced by Actipro CodeHighlighter (freeware)
http://www.CodeHighlighter.com/
-->
ORACLE_HOME
=
"
/
oracle
/
crs" sl_tableList
=
{"rac01:rac01
-
priv:rac01
-
vip:N:Y","rac02:rac02
-
priv:rac02
-
vip:N:Y"} ret_PrivIntrList
=
{"e1000g0:
10.198
.
88.0
:
1
","e1000g1:
192.168
.
0.0
:
2
"} n_storageTypeOCR
=
2
s_ocrpartitionlocation
=
"
/
oracle
/
ocrfile1" s_ocrMirrorLocation
=
"" n_storageTypeVDSK
=
2
s_votingdisklocation
=
"
/
oracle
/
vdfile1" s_OcrVdskMirror1RetVal
=
"" s_VdskMirror2RetVal
=
"" ORACLE_HOME_NAME
=
"OraCRS10ghome1" ################# complete modify ##################### RESPONSEFILE_VERSION
=
2.2
.
1.0
.
0
UNIX_GROUP_NAME
=
"oinstall" FROM_LOCATION
=
"..
/
stage
/
products.xml" NEXT_SESSION_RESPONSE
=<
Value Unspecified
>
TOPLEVEL_COMPONENT
=
{"oracle.crs","
10.2
.
0.1
.
0
"} DEINSTALL_LIST
=
{"oracle.crs","
10.2
.
0.1
.
0
"} SHOW_SPLASH_SCREEN
=
false SHOW_WELCOME_PAGE
=
false SHOW_NODE_SELECTION_PAGE
=
false SHOW_SUMMARY_PAGE
=
false SHOW_INSTALL_PROGRESS_PAGE
=
false SHOW_CONFIG_TOOL_PAGE
=
false SHOW_XML_PREREQ_PAGE
=
false SHOW_ROOTSH_CONFIRMATION
=
true SHOW_END_SESSION_PAGE
=
false SHOW_EXIT_CONFIRMATION
=
false NEXT_SESSION
=
false NEXT_SESSION_ON_FAIL
=
false SHOW_DEINSTALL_CONFIRMATION
=
false SHOW_DEINSTALL_PROGRESS
=
false RESTART_SYSTEM
=
false RESTART_REMOTE_SYSTEM
=
false REMOVE_HOMES
=
ORACLE_HOSTNAME
=<
Value Unspecified
>
SHOW_END_OF_INSTALL_MSGS
=
false COMPONENT_LANGUAGES
=
{"en"} s_clustername
=
"crs" CLUSTER_CONFIGURATION_FILE
=
""
8
. Oracle数据库软件的安装响应文件db.rsp，其中在####之前需要修改 CLUSTER_NODES
=
{} ORACLE_HOME_NAME
=
"OraDB10ghome1" ORACLE_HOME
=
"
/
oracle
/
10g" ########################################################## FROM_LOCATION
=
"..
/
stage
/
products.xml" RESPONSEFILE_VERSION
=
2.2
.
1.0
.
0
UNIX_GROUP_NAME
=
"oinstall" FROM_LOCATION_CD_LABEL
=<
Value Unspecified
>
SHOW_WELCOME_PAGE
=
true SHOW_CUSTOM_TREE_PAGE
=
true SHOW_COMPONENT_LOCATIONS_PAGE
=
true SHOW_SUMMARY_PAGE
=
true SHOW_INSTALL_PROGRESS_PAGE
=
true SHOW_REQUIRED_CONFIG_TOOL_PAGE
=
true SHOW_CONFIG_TOOL_PAGE
=
true SHOW_RELEASE_NOTES
=
true SHOW_ROOTSH_CONFIRMATION
=
true SHOW_END_SESSION_PAGE
=
true SHOW_EXIT_CONFIRMATION
=
true NEXT_SESSION
=
false NEXT_SESSION_ON_FAIL
=
true NEXT_SESSION_RESPONSE
=<
Value Unspecified
>
DEINSTALL_LIST
=
{"oracle.server","
10.2
.
0.1
.
0
"} SHOW_DEINSTALL_CONFIRMATION
=
true SHOW_DEINSTALL_PROGRESS
=
true ACCEPT_LICENSE_AGREEMENT
=
false TOPLEVEL_COMPONENT
=
{"oracle.server","
10.2
.
0.1
.
0
"} SHOW_SPLASH_SCREEN
=
true SELECTED_LANGUAGES
=
{"en"} COMPONENT_LANGUAGES
=
{"en"} INSTALL_TYPE
=
"Enterprise Edition" sl_superAdminPasswds
=<
Value Unspecified
>
sl_dlgASMCfgSelectableDisks
=
{} s_superAdminSamePasswd
=<
Value Unspecified
>
s_globalDBName
=
"orcl" s_dlgASMCfgRedundancyValue
=
"
2
(Norm)" s_dlgASMCfgNewDisksSize
=
"
0
" s_dlgASMCfgExistingFreeSpace
=
"
0
" s_dlgASMCfgDiskGroupName
=
"DATA" s_dlgASMCfgDiskDiscoveryString
=
"" s_dlgASMCfgAdditionalSpaceNeeded
=
" MB" s_dbSelectedUsesASM
=
"" s_dbSIDSelectedForUpgrade
=
"" s_dbRetChar
=
"" s_dbOHSelectedForUpgrade
=
"" s_ASMSYSPassword
=<
Value Unspecified
>
n_performUpgrade
=
0
n_dlgASMCfgRedundancySelected
=
2
n_dbType
=
1
n_dbSelection
=
0
b_useSamePassword
=
false b_useFileSystemForRecovery
=
true b_receiveEmailNotification
=
false b_loadExampleSchemas
=
false b_enableAutoBackup
=
false b_dlgASMShowCandidateDisks
=
true b_centrallyManageASMInstance
=
true sl_dlgASMDskGrpSelectedGroup
=
{" "," "," "," "} s_dlgRBOUsername
=
"" s_dlgEMCentralAgentSelected
=
"No Agents Found" b_useDBControl
=
true s_superAdminSamePasswdAgain
=<
Value Unspecified
>
s_dlgEMSMTPServer
=
"" s_dlgEMEmailAddress
=
"" s_dlgRBORecoveryLocation
=
"
/
oracle
/
db
/
flash_recovery_area
/
" n_upgradeDB
=
1
n_configurationOption
=
3c sl_upgradableSIDBInstances
=
{} n_upgradeASM
=
0
sl_dlgASMCfgDiskSelections
=
{} s_ASMSYSPasswordAgain
=<
Value Unspecified
>
n_dbStorageType
=
0
s_rawDeviceMapFileLocation
=
"" sl_upgradableRACDBInstances
=
{} s_dlgRBOPassword
=<
Value Unspecified
>
b_stateOfUpgradeDBCheckbox
=
false s_dbSid
=
"orcl" b_dbSelectedUsesASM
=
false sl_superAdminPasswdsAgain
=<
Value Unspecified
>
s_mountPoint
=
"
/
oracle
/
db
/
oradata
/
" b_stateOfUpgradeASMCheckbox
=
false oracle.assistants.server:OPTIONAL_CONFIG_TOOLS
=
"{}" oracle.has.common:OPTIONAL_CONFIG_TOOLS
=
"{}" oracle.network.client:OPTIONAL_CONFIG_TOOLS
=
"{}" oracle.sqlplus.isqlplus:OPTIONAL_CONFIG_TOOLS
=
"{}" oracle.sysman.console.db:OPTIONAL_CONFIG_TOOLS
=
"{}" varSelect
=
1
s_nameForOPERGrp
=
"dba" s_nameForDBAGrp
=
"dba"
