This appendix discusses the following topics:
Some messages recommend calling Oracle to report a problem. When you call your Oracle Support representative, have the following information available:
The hardware, operating system, and release number of the operating system on which the Oracle software is running
The complete release number of Oracle and other product software
All Oracle programs (with release numbers) in use when the error occurred
If you encountered one or more error codes or messages, then have the exact code numbers and message texts, in the order that they were displayed
Provide the exact text of Oracle Fail Safe messages (if any) that were written to the Windows Application Event Log
The problem severity, according to the following codes:
Your personal and company information:
Name
Company name
Company Oracle Support ID Number
Phone number
In some cases, Oracle Support Services will request a trace file.
See Section A.4 for information about using the trace function to log error output to a file.
To find the version of software that you run in the Oracle Fail Safe Manager help menu, select Help in the menu bar, then select About Oracle Fail Safe Manager. Version information for Oracle products that are integrated with Oracle Fail Safe is displayed in the output window for the Verify cluster command.
Oracle Fail Safe Manager error messages are saved in three ways. They are as follows:
Progress Window: This window displays the error messages to the user. Select Save As button to save the contents of the output window to a file that has more details, such as error numbers, timestamps, and version information.
Windows Application Event Log: Oracle Fail Safe resource monitor -- the cluster component that starts, stops, and monitors Oracle cluster resources -- posts error information in the Windows Application Event Log. Check that log if errors are encountered related to starting, stopping, or Is Alive polling of Oracle cluster resources.
Oracle Fail Safe trace files: Oracle Fail Safe logs more detailed information in these files that may provide clues to help determine the cause of errors.
Tracing is available to help you track, report, and examine errors that you receive in Oracle Fail Safe by dumping information about the errors to a log file.
Enable tracing for each node.
Follow these steps to enable tracing and set tracing flags on the cluster server nodes:
Run the Windows registry editor.
Select the following from the Registry tree:
HKEY_LOCAL_MACHINE, then SOFTWARE, then ORACLE, then FailSafe, and finally, Tracing
From the Registry Editor menu bar, select Edit, then select Add Value to open the Add String dialog box.
In the Value Name field, enter an Oracle Fail Safe value from Table A-1.
In the Data Type field, enter REG_SZ.
Click OK to open the String Editor dialog box.
In the String field, enter one or more of the Oracle Fail Safe strings shown in Table A-1. Separate multiple entries with commas.
Repeat steps 3 through 7 to set additional Oracle Fail Safe trace flags.
Restart the Cluster Service service on each node where tracing will be enabled. Note that, stopping the Cluster Service service will cause all cluster resources to fail over to another node in the cluster.
Table A-1 Trace Flags for Cluster Server Nodes
Note:
Oracle recommends usingALL
for FSS_TRACE_FLAGS
.Oracle Fail Safe trace files must be directed to a private disk.
Database trace and alert files can be located on either a cluster disk or a private disk:
If you use a cluster disk, then trace and alert files contain complete information about the operation. However, information about the node hosting the database is not recorded. The cluster disk used for these files must be one of the disks used for the archive log files or the database data files (where Create Sample Database places them, for example); otherwise, they will not be added to the group.
If you use a private disk, then trace and alert files each contain node-specific information about the operation. However, you must view files from each cluster node at the same time to obtain complete chronological information if the database has failed over or been moved. Use a path name that is valid on each node so that data can be written to these files correctly. Files on private disks are never added to a group.