PRIMECLUSTER Wizard for Oracle Configuration and Administration Guide 4.2 - Linux for Itanium -
Contents PreviousNext

Appendix A Failure case studies> A.1 Oracle instances

A.1.1 Startup and stop failure

Oracle instance startup or stop failed.

[Case 1] (Standby, Oracle RAC Scalable)

Oracle instance startup or stop fails when the PRIMECLUSTER RMS resource settings are invalid.

Check the resource setup information by executing the "clorainfo -c" command. Refer to "4.6 clorainfo - Display Resource Configuration and Monitoring Status".

If your setup information is incorrect, set proper values again. Refer to "Chapter 2 Environment Setup".

[Case 2] (Standby)

Oracle instance startup or stop fails when user script settings in /opt/FJSVclora/usr are incorrect.

Check if the user scripts behave correctly.

[Case 3] (Standby, Oracle RAC Scalable)

Oracle instance startup or stop fails when Oracle settings are invalid.

Check the Oracle alertlog, so you might be able to detect the cause of the failure.
Also, check if Oracle can be started or stopped properly. If you discontinue Oracle instance resource monitoring, you can operate Oracle manually. Refer to "3.2.4.1 Oracle must be stopped".

[Case 4] (Oracle RAC Scalable)

When starting userApplication with Oracle RAC Instances resource in the state where Oracle Clusterware is not running, the resource state enters Wait until exceeds the time set as ScriptTimeout (refer to step "12" - 2.3.5). If it exceeds the time set as ScriptTimeout with the resource state maintain Wait, the resource state will enter Faulted.

Execute the following methods according to each pattern.

- In the case of configuration that the Oracle Clusterware and Oracle RAC instance are registered in different userApplication. (It matches "Configuration A" in "2.3.3 userApplication Creation".)

At first, check whether Oracle Clusterware is running.

The following command execution, you can check whether Oracle Clusterware is running.

# su - <Oracle user>
$ crs_stat


When the following message is outputted, Oracle Clusterware may be not running.

CRS-0184: Cannot communicate with the CRS daemon.

When Oracle Clusterware is not running, there is a possibility that the setting that Oracle Clusterware starts earlier than Oracle RAC Instance is incorrect.

Check whether the script which starts Oracle Clusterware resource is set as the PreOnlineScript of Oracle RAC Instance (refer to "note" of step "8" of "2.3.5 userApplication with Oracle RAC Instances and Listeners").

- In the case of configuration that the Oracle Clusterware and Oracle RAC instance are registered in the same userApplication. (It matches "Configuration B" in "2.3.3 userApplication Creation".)

There is a possibility that starting of Oracle RAC Instance resource is failed.

Check whether you have no mistakes in setting of Oracle RAC Instance resource (2.3.5 userApplication with Oracle RAC Instances and Listeners).

In addition, check whether you have no mistakes in setting of PRIMECLUSTER RMS and Oracle (refer to "Case 1" and "Case 3").

- In the case of configuration that the Oracle Clusterware is not registered. (It matches "Configuration C" in "2.3.3 userApplication Creation".)

Check whether Oracle Clusterware is running.

The following command execution, you can check whether Oracle Clusterware is running.

# su - <Oracle user>
$ crs_stat

When the following messages are outputted, Oracle Clusterware may be not running.

CRS-0184: Cannot communicate with the CRS daemon.

Clear Faulted state of the resource, after that, execute the following commands and start Oracle Clusterware.

# /etc/init.d/init.crs start


After that, start userApplication.


Contents PreviousNext

All Rights Reserved, Copyright(C) FUJITSU LIMITED 2009