配置对不对? from metalink;Cluster Manager configuration files on Oracle9.2.0.4 on Linux -------------------------------------------------------------
Most of the cluster Manager errors are due to incorrect configuration files. Please ensure you have the exactly the same cmcfg.ora file across all nodes. Sample of the file ::-
[oracle@skacker-linux admin]$ more cmcfg.ora HeartBeat=15000 ClusterName=Oracle Cluster Manager, version 9i PollInterval=1000 MissCount=210 PrivateNodeNames=skacker-linux skacker-linux2 PublicNodeNames=skacker-linux skacker-linux2 ServicePort=9998 CmDiskFile=/dev/raw/raw14 <<< make sure this is same on all nodes HostName=skacker-linux <<< this should be only change in this file across all nodes KernelModuleName=hangcheck-timer [oracle@skacker-linux admin]$
Also ensure that the Watchdog demon is turned off and nodes is reboot. Comment the line as shown below in file ocmargs.ora ::-
It will help to create a file, say ocm, below as execute it as root to start cluster manager
[root@skacker-linux root]# more ocm #!/bin/bash PATH=/opt/oracle/OraHome1/bin:$PATH export PATH ORACLE_BASE=/opt/oracle/OraHome1 ORACLE_HOME=/opt/oracle/OraHome1 export ORACLE_HOME export ORACLE_BASE rm -rf $ORACLE_HOME/oracm/log/*.ts $ORACLE_HOME/oracm/bin/ocmstart.shFinally, pre-create the quorum file as a larger size with a dd command:dd if=/dev/zero of=<quorum disk used for CM> bs=4096 count=65 ( you can use 100 also )This will create the quorum file as 266240 bytes (bs=4096 * count=65 = 266240)
from metalink;Cluster Manager configuration files on Oracle9.2.0.4 on Linux
-------------------------------------------------------------
Most of the cluster Manager errors are due to incorrect configuration files. Please ensure you have the exactly the same cmcfg.ora file across all nodes.
Sample of the file ::-
[oracle@skacker-linux admin]$ more cmcfg.ora
HeartBeat=15000
ClusterName=Oracle Cluster Manager, version 9i
PollInterval=1000
MissCount=210
PrivateNodeNames=skacker-linux skacker-linux2
PublicNodeNames=skacker-linux skacker-linux2
ServicePort=9998
CmDiskFile=/dev/raw/raw14 <<< make sure this is same on all nodes
HostName=skacker-linux <<< this should be only change in this file across all nodes
KernelModuleName=hangcheck-timer
[oracle@skacker-linux admin]$
Also ensure that the Watchdog demon is turned off and nodes is reboot. Comment the line as shown below in file ocmargs.ora ::-
[oracle@skacker-linux admin]$ more ocmargs.ora
# Sample configuration file $ORACLE_HOME/oracm/admin/ocmargs.ora
#watchdogd -d /dev/null -l 0 -m 0
oracm /a:0
norestart 1800
[oracle@skacker-linux admin]$
It will help to create a file, say ocm, below as execute it as root to start cluster manager
[root@skacker-linux root]# more ocm
#!/bin/bash
PATH=/opt/oracle/OraHome1/bin:$PATH
export PATH
ORACLE_BASE=/opt/oracle/OraHome1
ORACLE_HOME=/opt/oracle/OraHome1
export ORACLE_HOME
export ORACLE_BASE
rm -rf $ORACLE_HOME/oracm/log/*.ts
$ORACLE_HOME/oracm/bin/ocmstart.shFinally, pre-create the quorum file as a larger size with a dd command:dd if=/dev/zero of=<quorum disk used for CM> bs=4096 count=65 ( you can use 100 also )This will create the quorum file as 266240 bytes (bs=4096 * count=65 = 266240)