If the WebLogic domain folder that was specified during the Administrative UI installation is not located under the default WebLogic installation folder, the installer may not use the specified folder. Rather, the installer may use an incorrect domain folder.
Symptom:
The Oracle RAC nodes propagate changes within 7 seconds. SiteMinder could read and write objects to a policy store, user store, session store, or audit store more often. As a result, the default Oracle RAC propagation window can result in SiteMinder errors. These SiteMinder errors occur because the write operation was made into one node and the read operation was made to another node.
Solution:
Configure the following setting in the Oracle RAC cluster:
MAX_COMMIT_PROPAGATION_DELAY=0
Note: For more information about configuring this setting, see the Oracle documentation.
Symptom:
Under heavy load, the Policy Server may fail to insert queued audit events into the audit store. If the failure occurs, the SiteMinder Policy Server log (smps.log) displays the following error:
[INFO] Failed attempt to bulk insert audit message: Code: -1044. DB Code: 2
Solution:
Two registry keys determine when the Policy Server inserts audit events into the audit database: SQLBulkInsertFlushInterval and SQLBulkInsertFlushRowCount:
Modify the SQLBulkInsertFlushRowCount registry key to resolve the error message.
To modify the registry key
Specifies the Policy Server installation path.
Increase the value to be at least twice as large as the number of audit events that were created, per second, when the error appeared in the SiteMinder Policy Server log.
Example: If 1,500 audit events occurred when the error appeared, increase the value to 3,000.
Copyright © 2012 CA.
All rights reserved.
|
|