类 CommandLineJobRunner

Basic launcher for starting jobs from the command line. In general, it is assumed that this launcher will primarily be used to start a job via a script from an Enterprise Scheduler. Therefore, exit codes are mapped to integers so that schedulers can use the returned values to determine the next course of action. The returned values can also be useful to operations teams in determining what should happen upon failure. For example, a returned code of 5 might mean that some resource wasn't available and the job should be restarted. However, a code of 10 might mean that something critical has happened and the issue should be escalated.

With any launch of a batch job within Spring Batch, a Spring context containing the Job and some execution context has to be created. This command line launcher can be used to load the job and its context from a single location. All dependencies of the launcher will then be satisfied by autowiring by type from the combined application context. Default values are provided for all fields except the JobLauncher and JobLocator . Therefore, if autowiring fails to set it (it should be noted that dependency checking is disabled because most of the fields have default values and thus don't require dependencies to be fulfilled via autowiring) then an exception will be thrown. It should also be noted that even if an exception is thrown by this class, it will be mapped to an integer and returned.

Notice a property is available to set the SystemExiter. This class is used to exit from the main method, rather than calling System.exit() directly. This is because unit testing a class the calls System.exit() is impossible without kicking off the test within a new JVM, which it is possible to do, however it is a complex solution, much more so than strategizing the exiter.

The arguments to this class can be provided on the command line (separated by spaces), or through stdin (separated by new line). They are as follows:

The command line options are as follows

• jobPath: the xml application context containing a Job
• -restart: (optional) to restart the last failed execution
• -stop: (optional) to stop a running execution
• -abandon: (optional) to abandon a stopped execution
• -next: (optional) to start the next in a sequence according to the JobParametersIncrementer in the Job
• jobIdentifier: the name of the job or the id of a job execution (for -stop, -abandon or -restart).
• jobParameters: 0 to many parameters that will be used to launch a job specified in the form of key=value pairs.

If the -next option is used the parameters on the command line (if any) are appended to those retrieved from the incrementer, overriding any with the same key.

The combined application context must contain only one instance of JobLauncher. The job parameters passed in to the command line will be converted to Properties by assuming that each individual element is one parameter that is separated by an equals sign. For example, "vendor.id=290232". The resulting properties instance is converted to JobParameters using a JobParametersConverter from the application context (if there is one, or a DefaultJobParametersConverter otherwise). Below is an example arguments list: "

java org.springframework.batch.core.launch.support.CommandLineJobRunner testJob.xml testJob schedule.date=2008/01/24 vendor.id=3902483920

Once arguments have been successfully parsed, autowiring will be used to set various dependencies. The JobLauncher for example, will be loaded this way. If none is contained in the bean factory (it searches by type) then a BeanDefinitionStoreException will be thrown. The same exception will also be thrown if there is more than one present. Assuming the JobLauncher has been set correctly, the jobIdentifier argument will be used to obtain an actual Job. If a JobLocator has been set, then it will be used, if not the beanFactory will be asked, using the jobIdentifier as the bean id.