Spring Batch_JOB执行流程分析java
基于如下job的配置,spring
<batch:job id="addPeopleDescJob"> <batch:step id="addDescStep" parent="abstractStep"> <batch:tasklet> <batch:chunk reader="peopleAddDescReader" processor="allowSkipProcessor" writer="addDescPeopleWriter" commit-interval="2"> </batch:chunk> </batch:tasklet> </batch:step> </batch:job>
debug代码ide
JobExecution result = launcher.run(job, jobParametersBuilder.toJobParameters());
这是启动job的方法,以下是方法的具体实现:post
SimpleJobLauncher.java run方法的具体实现(删除了部分代码)ui
@Override public JobExecution run(final Job job, final JobParameters jobParameters) throws JobExecutionAlreadyRunningException, JobRestartException, JobInstanceAlreadyCompleteException, JobParametersInvalidException { final JobExecution jobExecution; jobExecution = jobRepository.createJobExecution(job.getName(), jobParameters); try { taskExecutor.execute(new Runnable() { @Override public void run() { try { job.execute(jobExecution); } catch (Throwable t) { rethrow(t); } } private void rethrow(Throwable t) { } }); } return jobExecution; }
看taskExecutor.execute 方法的主要逻辑,就是把 job execute 封装进一个抽象的任务内,经过taskExecutor 执行 ,taskExecutor 是一个什么类型——org.springframework.core.task.SyncTaskExecutor,一个同步的任务执行类。这样就能够明确的知道每一个chunk的处理都是在一个单线程内,循环往复的处理每一个commit-interval。this
taskExecutor.execute(new Runnable() { @Override public void run() { job.execute(jobExecution); } });
继续走代码,那接下来执行的方法是job.execute(jobExecution);那先来看job 是什么类型的对象,spa
那么就是他了,FlowJob: [name=addPeopleDescJob]。进入job.execute ,线程
他就直接跳进了 org.springframework.batch.core.job.AbstractJob.execute(JobExecution execution)方法,这是FlowJob的父类,FlowJob 显然没有覆写该方法。debug
AbstractJob.execute(JobExecution execution)code
@Override public final void execute(JobExecution execution) { doExecute(execution); }
这是方法的主要执行逻辑,那么这个doExecute方法就是FlowJob实现父类的抽象方法,完成 job的执行的任务。
FlowJob.doExecute()
/** * @see AbstractJob#doExecute(JobExecution) */ @Override protected void doExecute(final JobExecution execution) throws JobExecutionException { try { JobFlowExecutor executor = new JobFlowExecutor(getJobRepository(), new SimpleStepHandler(getJobRepository()), execution); executor.updateJobExecutionStatus(flow.start(executor).getStatus()); } catch (FlowExecutionException e) { if (e.getCause() instanceof JobExecutionException) { throw (JobExecutionException) e.getCause(); } throw new JobExecutionException("Flow execution ended unexpectedly", e); } }
>>>>>>>>>>>>>>>>>>>>>>>>>>>
继续debug
job是任务的任务的抽象表示,完成的具体任务还要在step中,那么接下来就是step的执行了,step是如何执行的?
咱们的代码最终调到这org.springframework.batch.core.step.AbstractStep.execute (StepExecution stepExecution) ,step的抽象类。这个方法的主要逻辑以下:
@Override public final void execute(StepExecution stepExecution) throws JobInterruptedException, UnexpectedJobExecutionException { doExecute(stepExecution); }
这个step又是什么类型的对象——TaskletStep: [name=addDescStep],其父类为AbstractStep。
继续看TaskletStep.doExecute(stepExecution);
@Override protected void doExecute(StepExecution stepExecution) throws Exception { stepExecution.getExecutionContext().put(TASKLET_TYPE_KEY, tasklet.getClass().getName()); stepExecution.getExecutionContext().put(STEP_TYPE_KEY, this.getClass().getName()); stream.update(stepExecution.getExecutionContext()); getJobRepository().updateExecutionContext(stepExecution); // Shared semaphore per step execution, so other step executions can run // in parallel without needing the lock final Semaphore semaphore = createSemaphore(); stepOperations.iterate(new StepContextRepeatCallback(stepExecution) { @Override public RepeatStatus doInChunkContext(RepeatContext repeatContext, ChunkContext chunkContext) throws Exception { StepExecution stepExecution = chunkContext.getStepContext().getStepExecution(); // Before starting a new transaction, check for // interruption. interruptionPolicy.checkInterrupted(stepExecution); RepeatStatus result; try { result = new TransactionTemplate(transactionManager, transactionAttribute) .execute(new ChunkTransactionCallback(chunkContext, semaphore)); } catch (UncheckedTransactionException e) { // Allow checked exceptions to be thrown inside callback throw (Exception) e.getCause(); } chunkListener.afterChunk(chunkContext); // Check for interruption after transaction as well, so that // the interrupted exception is correctly propagated up to // caller interruptionPolicy.checkInterrupted(stepExecution); return result; } }); }
上面就是TaskletStep.doExecute的完整代码,其中主要的逻辑是:
stepOperations.iterate(new StepContextRepeatCallback(stepExecution) { @Override public RepeatStatus doInChunkContext(RepeatContext repeatContext, ChunkContext chunkContext) throws Exception { RepeatStatus result; try { result = new TransactionTemplate(transactionManager, transactionAttribute) .execute(new ChunkTransactionCallback(chunkContext, semaphore)); } return result; } });
stepOperations.iterate() 方法传入StepContextRepeatCallback的一个匿名对象。
继续debug,转了一圈又回到这个回调方法上StepContextRepeatCallback.doInChunkContext(),
result = new TransactionTemplate(transactionManager, transactionAttribute) .execute(new ChunkTransactionCallback(chunkContext, semaphore));
new TransactionTemplate().execute()方法忽略过去,在这个方法里面最终还要调用ChunkTransactionCallback.doInTransaction() 的回调方法:
@Override public RepeatStatus doInTransaction(TransactionStatus status) { RepeatStatus result = RepeatStatus.CONTINUABLE; result = tasklet.execute(contribution, chunkContext); return result; }
这个tasklet 的类型是 :org.springframework.batch.core.step.item.ChunkOrientedTasklet ,定义step的执行策略。ChunkOrientedTasklet.execute() 方法:
@Override public RepeatStatus execute(StepContribution contribution, ChunkContext chunkContext) throws Exception { @SuppressWarnings("unchecked") Chunk<I> inputs = (Chunk<I>) chunkContext.getAttribute(INPUTS_KEY); if (inputs == null) { inputs = chunkProvider.provide(contribution); if (buffering) { chunkContext.setAttribute(INPUTS_KEY, inputs); } } chunkProcessor.process(contribution, inputs); chunkProvider.postProcess(contribution, inputs); // Allow a message coming back from the processor to say that we // are not done yet if (inputs.isBusy()) { logger.debug("Inputs still busy"); return RepeatStatus.CONTINUABLE; } chunkContext.removeAttribute(INPUTS_KEY); chunkContext.setComplete(); logger.debug("Inputs not busy, ended: " + inputs.isEnd()); return RepeatStatus.continueIf(!inputs.isEnd()); }
Chunk<I> inputs,就是要读入的数据,他是怎么来的 :
inputs = chunkProvider.provide(contribution);
chunkProvider 是 org.springframework.batch.core.step.item.SimpleChunkProvider 的类型,provide方法:
@Override public Chunk<I> provide(final StepContribution contribution) throws Exception { final Chunk<I> inputs = new Chunk<I>(); repeatOperations.iterate(new RepeatCallback() { @Override public RepeatStatus doInIteration(final RepeatContext context) throws Exception { I item = null; try { item = read(contribution, inputs); } catch (SkipOverflowException e) { // read() tells us about an excess of skips by throwing an // exception return RepeatStatus.FINISHED; } if (item == null) { inputs.setEnd(); return RepeatStatus.FINISHED; } inputs.add(item); contribution.incrementReadCount(); return RepeatStatus.CONTINUABLE; } }); return inputs; }
在 provide 方法内,逐条读取一块数据(经过jdbc游标来读取一条或经过分页来读取多条)item = read(contribution, inputs);而后放入inputs 数据块chunk inputs 。数据读完了(此commit内,根据commit-interval的配置,读取配置的次数或方法返回null表示数据读取完成)接下来就要处理inputs 数据了,到chunkProcessor.process (contribution, inputs) :
@Override public final void process(StepContribution contribution, Chunk<I> inputs) throws Exception { // Allow temporary state to be stored in the user data field initializeUserData(inputs); // If there is no input we don't have to do anything more if (isComplete(inputs)) { return; } // Make the transformation, calling remove() on the inputs iterator if // any items are filtered. Might throw exception and cause rollback. Chunk<O> outputs = transform(contribution, inputs); // Adjust the filter count based on available data contribution.incrementFilterCount(getFilterCount(inputs, outputs)); // Adjust the outputs if necessary for housekeeping purposes, and then // write them out... write(contribution, inputs, getAdjustedOutputs(inputs, outputs)); }
在chunkProcessor.process 方法内定义了 两个步骤,一个是transform,也就是process读入的数据集合,一个是write。代码跳到transform这:
protected Chunk<O> transform(StepContribution contribution, Chunk<I> inputs) throws Exception { Chunk<O> outputs = new Chunk<O>(); for (Chunk<I>.ChunkIterator iterator = inputs.iterator(); iterator.hasNext();) { final I item = iterator.next(); O output; try { output = doProcess(item); } catch (Exception e) { /* * For a simple chunk processor (no fault tolerance) we are done * here, so prevent any more processing of these inputs. */ inputs.clear(); throw e; } if (output != null) { outputs.add(output); } else { iterator.remove(); } } return outputs; }
对于 inputs 集合 ,遍历处理后获得 outputs 集合。
output = doProcess(item);
上面这行代码就是要调用自定义的process 方法进行处理。而后看write 方法是如何写入数据的,
protected void write(StepContribution contribution, Chunk<I> inputs, Chunk<O> outputs) throws Exception { try { doWrite(outputs.getItems()); } catch (Exception e) { /* * For a simple chunk processor (no fault tolerance) we are done * here, so prevent any more processing of these inputs. */ inputs.clear(); throw e; } contribution.incrementWriteCount(outputs.size()); }
在write 方法内调用了 doWrite() 方法 ,而后调用自定义的 write方法把数据写入。
上面就是 在一次 commit-interval 内的主要过程和主要的逻辑代码。那么 spring batch 是如何重复commit-interval 的呢 (经过RepeatTemplate)?之后再详细说来。
在course reader中,经过游标逐条的读取数据,造成一个inputs chunk,至关于一个集合,而后逐条的处理inputs chunk,进行item process,造成一个outpus chunk,由outputs chunk获得一个集合outputs.getItems(),直接处理这个集合,也就是write方法的参数是一个集合类型。
================END================