Release 2 (9.0.2)
Part Number A90500-02
 Home |
 Solution Area |
 Contents |
 Index |
| Oracle9iAS Clickstream Intelligence Administrator's Guide Release 2 (9.0.2) Part Number A90500-02 |
|
This chapter describes how to execute, monitor, and manage operations related to the Clickstream database. All functionality discussed in this chapter is contained within the Manage tab located on the right side of the Horizontal Navigation bar.
The items listed below represent the links that appear and become active on the Horizontal Navigation bar when you select the Manage tab. In this chapter, you'll learn about:
The Database Processes section of this chapter describes how to start, manage, and interpret details about the database processes you execute. This section covers the following topics:
The Process Status page is displayed by default when you click the Processes link on the Horizontal Navigation bar. It provides information related to the activity of your database, and is the page from which you can start a database process.
The Process Status page contains the following two headings:
The status of the Execution Engine is also displayed on the Process Status page. The Execution Engine should run continually in the background; it must be running in order to execute any Clickstream processes. If the Execution Engine is not currently running, a Warning message is displayed at the top of the Process Status page. (No message appears when the Engine is currently running.)
To start or stop the Execution Engine, use the same commands that also apply to the Collector Server. At the command prompt, go to the appropriate directory below.
(UNIX) CLICK_HOME/click/bin (Windows) CLICK_HOME\click\bin
In the expression above, CLICK_HOME is the directory in which Oracle9iAS Clickstream Intelligence is installed. Then, use the start and stop commands below.
clkctl start DATABASE_LOGIN clkctl stop DATABASE_LOGIN
In the expressions above, DATABASE_LOGIN is equal to the username for the Runtime Administrator schema, which should be entered as clkrt.
To connect to a Clickstream database instance other than the one specified in the click-app.xml file, the following commands must be used with the appropriate variables.
clkctl start DATABASE_LOGIN clkctl stop DATABASE_LOGIN
In the commands above, DATABASE_LOGIN is equal to username/password@HOST:PORT:SID The username entered must be clkrt, the password corresponds to the one defined for the clkrt schema, the HOST is the name of the machine on which the database is installed, the PORT is the port number of the database TNS listener, and SID corresponds to the Oracle SID.
To abort the Execution Engine (and Collector Server), enter the following command:
clkctl abort DATABASE_LOGIN
where DATABASE_LOGIN uses the appropriate variables described above. This command stops the daemons when a process is hanging or simply taking too long to complete. After using the abort command, you must run execln.sql from the appropriate directory below:
(UNIX) CLICK_HOME/admin (Windows) CLICK_HOME\admin
To view the status of the Clickstream daemons, use the command below:
clkctl status DATABASE_LOGIN
where DATABASE_LOGIN equals the appropriate variables described above.
The Start Process page enables you to select the database process you want to run. It lists the five Clickstream process types, accompanied by a pull-down menu that lists all existing definitions for that process type. If at least one definition does not appear for a particular process, you must first Create a Process Definition. (See Chapter 3, "Configuring Clickstream Intelligence".)
Only one database process can be run at a time. If a process is already running, you must either wait until the process has finished, or stop the currently running process. You can determine if a process is already in progress by checking the "Currently Running" section of the Process Status page. If a process is not already underway, then a new database process can be started.
To start a database process, follow the steps below:
The Start Process page appears.
The process is started by the Execution Engine. The Process Status page displays the process status as "Running" under the Currently Running heading.
To stop a currently running process, click the Stop button on the Process Status page. Because the process actually stops only after all currently running jobs are executed, the process may appear to remain in the Stopping state for an extended period of time.
When the Stopped execution state is finally reached, however, the process is still not entirely finished- you must click the Undo button to completely terminate the Stopped process.
The Previous Runs section of the Process Status page displays all completed and failed database processes. To view the details for a previously-run process definition, follow the steps below:
The Process Status page appears (by default).
The Process Details page displays information about the process, such as process type, definition, start date, status, and warehouse version.
You can define definitions for five types of database processes (see "Create a Process Definition" in Chapter 3.) To start a process, use the pull-down menu to select a particular process definition for one of the five available process types. The sections that follow describe each process type.
The Load Clickstream process enables the transfer and storage of Web log data into the Clickstream database. When you start a Load Clickstream process, the Clickstream Loader begins processing and transforming Web log data and then loads it into the database.
|
Note: When a Load Clickstream process is executed, the interface tables are first truncated (emptied) before Web log data is loaded into the Clickstream Intelligence database. Therefore, if you have loaded external (non-Web) data into the interface tables, a Load Dimensions process must be run immediately thereafter to ensure that interface table data is loaded into the database dimensions. If external data is loaded into the interface tables and then a Load Clickstream process (instead of a Load Dimensions process) is executed, all interface table data will be lost when the interface tables are emptied. |
The Load Dimensions process transfers existing interface table data into the database levels, which are used to populate the dimensions. This process does not load data into the interface tables - the user must load external or non-Web data into the interface tables before running the Load Dimensions process.
Summary Refresh updates the Summary Layer with the most current version of data in the database. It may be useful to note that you can automatically refresh summaries as part of the Load Clickstream process when the "Refresh Summaries" option is enabled. For more information, see the "Refresh Summaries" section of Chapter 3, "Configuring Clickstream Intelligence".
The Resolve Unknown IP Address process operates only on the Client Host dimension. This process is typically performed when the Web log contains the IP address of the host, but does not include the resolved name. To resolve the IP address, the Resolve Unknown IP Addresses process queries the DNS server (in a reverse DNS lookup) to determine the user-friendly hostname from the client's numerical IP address.
The Restore a Previous Version process enables you to undo any changes to data that have been made since a given point in time. The database version number to which you want to roll back is specified in the "Version after Rollback" field when you initially create a process definition of this type. For example, if you run a definition that indicates "2" as the database version after roll back, running the Restore a Previous Version process will return the database to the version previously labeled "2."
Status information about currently running processes is useful from both a tracking and a troubleshooting standpoint. For a process that is "Currently Running" on the Process Status page, you can click the Show Details button to view and track general process information via the Process Details page.
When a running process displays a status indicative of a problem (such as Error or Failed status), you can troubleshoot the problem by viewing Process Details and other information available from The Process Navigation Menu. Troubleshooting typically involves viewing Process Messages and analyzing details about Subprocesses (and the jobs of which they are comprised.)
Consider the following scenario in which a Load Clickstream process is running. A problem arises that causes the process to stop. The status of this process changes from Running to Error on the Process Status page, and three buttons are displayed:
This page is the first option on the Process Navigation menu and provides general information about the process, such as its name and definition. You can use the other options on The Process Navigation Menu to further investigate the process error (as described below.)
The process is listed under the "Previous Runs" heading on the Process Status page with the status Failed. To view process information and access the Process Navigation menu, click the Show Details button.
Depending on the option selected above, one of several execution states may ultimately result. Completed or Failed status indicates that execution for a given process has finished. An intermediate state, such as Error, indicates a process that has temporarily stopped, but must ultimately Resume or be cancelled via the Undo button.
The following sections outline typical process execution states and scenarios that may be encountered when running a database process.
If errors were not fixed - process reaches Error status again.
The Process Navigation menu is displayed whenever you view specific information about a process. It is used to view information about a particular process- its details, definition, messages, subprocesses and jobs, and data packets details.
The following options are accessible from the Process Navigation menu:
The following sections describe each item displayed on the Process Navigation menu.
The Process Details page provides additional information about a process. For a given process, the following information is displayed:
The Messages option on the Process Navigation menu provides information about the jobs that have been run for a process. The messages displayed typically indicate job errors, and can be used for troubleshooting a process that has failed or has an error.
The Process Messages page displays:
The Definition menu option displays the parameters (and values) that characterize the process. This information can be modified from the Process Definitions page, located under the Configure tab.
The Subprocesses page lists all subprocesses that were executed for a particular process. To view the jobs the comprise a particular subprocess, select the radio button beside the subprocess and click Show Details.
This page lists all jobs that comprise a given subprocess. For each job, additional statistics are displayed, such as the Job Type, Start Date, End Date, and Status. To drill down into the details of any given job, click the Show Details button.
The Job Details page provides the most granular data about a database process - details about a specific job that was executed in the database. The job name is displayed with the actual command that was used to start the execution of the job and the job messages (if any) that were generated. The information provided on the Job Details page is typically used for detailed analysis of database processes, perhaps when a process fails or has an error.
Data packets consist of Web log files that are grouped together and compressed by the Collector Agent (installed on the Web server). Formation of data packets facilitates transmission from the Web server to the Clickstream Collector Server.
The Process Data Packets Details page displays specific parameters associated with the data packets for a given file. Information such as File Name, Status, Date Created, Lines, Lines Loaded, Lines Rejected, and Lines Discarded may be displayed for any data packet associated with a given data source.
To delete data packets from the Process Data Packets page, select the radio button for the packets you want to permanently remove and click Delete. Deletion of data packets is permanent. Periodic deletion of data packets is recommended, as it creates free (unused) disk space on the Collector Server host.
The Data Sources link on the Horizontal Navigation bar provides access to the Data Sources main page. This page displays data sources for all sites in your system. To view all data packets associated with any source listed on the Data Sources page, select the appropriate radio button and click Show Data Packets.
The Data Packets main page lists, in tabular format, all data packets that have been downloaded by the Collector Server.
|
Note: All data packets are listed for a data source, even if the packets have not been loaded into the database. |
From the Data Packets page, you can:
When you select a data source from the Data Sources main page and click the Show Data Packets button, the Data Packets page appears with a listing of all data packets that have been downloaded for that specific data source.
To view details about an item listed on the Data Packets page, locate the data packets file name and click the corresponding Show Details link. The Data Packets Details page appears with the following information:
|
 Copyright © 2002 Oracle Corporation. All Rights Reserved. |
|
