Activating and configuring Log4j
Talend Studio includes the Apache logging utility Log4j to provide logging at runtime. You can enable or disable Log4j loggers in components and customize the Log4j configuration globally for the project.
To avoid any Job failure, clear the Activate log4j in components check box from the Log4j view in or clear the log4jLevel check box from the Advanced settings view of your Spark Job.
Procedure
-
On the toolbar of the Talend Studio
main window, click
or click from the menu bar to open the Project Settings
dialog box.
-
In the tree view of the dialog box, click the Log4j node to
open the Log4j view.
-
If needed, change the global Log4j configuration by modifying the XML instructions
in the Log4j template area.
For example, to configure the root logger for Log4j 2 to output all debug or higher messages, go to the Loggers section and set the value of the level attribute of the root node to debug.
The Log4j 2 MDC (Mapped Diagnostic Context) is populated with key-value pairs. These extra log attributes help you easily identify the thread in which log data is generated when all log messages from multiple threads are written into a single file. For more information about Log4j 2 MDC, see Log4j 2 API - Thread Context.
To include all extra log attributes when writing logs, add %X in the PatternLayout section.
<PatternLayout pattern="[%-5level] %d{HH:mm:ss} %logger{36}- %X %msg%n" />To include specific log attributes when writing logs, add %X{key} %X{key2} ... %X{keyN} in the PatternLayout section, where {keyN} is the key of the corresponding log attribute. You can also use characters like commas to separate attributes to easily identify them. The following example includes the _pid and _startTimestamp attributes when writing logs.
<PatternLayout pattern="[%-5level] %d{HH:mm:ss} %logger{36}- %X{_pid}, %X{_startTimestamp} %msg%n" />For more information about Log4j 2 configuration, see Configuring with XML.
For more information about Log4j 1 configuration, see Log4jXmlFormat.