When you first set up your high availability Lucene configuration, you can configure Novell Teaming to write Lucene node information to the Tomcat log file so that you can observer the behavior of the Lucene nodes. Then you can take nodes down and bring them up again to see the effect on your Teaming site. This process helps you prepare for the situation where a Lucene node goes down unexpectedly or you need to take one down on purpose, perhaps for maintenance.
NOTE:For background information about the Tomcat log file, see Tomcat Log File
in Site Maintenance
in the Novell Teaming 2.0 Administration Guide.
By default, Novell Teaming writes an error to the Tomcat log file whenever it cannot communicate with a Lucene node. For testing purposes, you can increase the logging so that the Tomcat log file includes a message for each time Teaming contacts a Lucene Index Server. After observing and understanding the behavior, you should return the logging to normal levels.
Open the log4j.properties file in a text editor.
The location of the log4j.properties file varies by platform.
Search for the following line:
#log4j.category.com.novell.teaming.search=DEBUG
Remove the pound sign (#) from the beginning of the line to activate DEBUG level logging for the Lucene Index Server.
Save the log4j.properties file, then exit the text editor.
Stop Teaming, then start Teaming to put the new logging level into effect.
After changing the Novell Teaming logging level to include Lucene node activity, you can monitor the Tomcat log file to see which Lucene nodes Teaming is contacting.
Change to the directory where the Tomcat log file is located.
For background information about the Tomcat log file, see Tomcat Log File
in Site Maintenance
in the Novell Teaming 2.0 Administration Guide.
On Linux, use the tail command to monitor the end of the Tomcat log file.
tail -f catalina.out
or
On Windows, download an equivalent command from the Internet in order to perform the following steps.
For example, there is one available at SourceForge.net.
Perform some activities on the Teaming site to create new content.
For example, you could create a blog entry. The Tomcat log file shows that Teaming is contacting each available Lucene Index Server in turn as it submits the new information for indexing. Nodes are considered available if they are marked in the box on the Lucene Nodes page.
Bring down one of the Lucene nodes.
For example, you might reboot the Lucene server, stop the Lucene Index Server on the node, or otherwise prevent Teaming from contacting the Lucene node.
Observe that the Tomcat log file displays an error when the Lucene node becomes unavailable.
Perform some additional activities on the Teaming site that create new content.
You might notice a pause as Teaming tries to connect with the Lucene node that is no longer available.
Observe that the Tomcat log file shows when normal Teaming processing has resumed without access to the unavailable Lucene node.
Make the unavailable Lucene node available again.
Observe that the Tomcat log file indicates that Teaming has created a journal record on the Lucene node that was temporarily unavailable.
The journal record contains the indexing operations that failed to take place while the Lucene node was unavailable. Even though the Lucene node is back up, Teaming does not start accessing it until it has been synchronized with the latest index information.
To synchronize the out-of-date Lucene node, follow the steps in Section 16.4, Synchronizing a High Availability Lucene Configuration.
After the out-of-date Lucene node has been synchronized, Teaming starts accessing it again.
If a situation arises where no Lucene node is marked in the box on the Lucene Nodes page, Teaming temporarily accesses the first node in the list that is set . If no nodes are set to or , Teaming temporarily access the first node in the list, even though it is set to . This functionality is required because you cannot log in to the Teaming site without access to a Lucene Index Server.