org.gridgain.grid.ggfs.mapreduce

Interface GridGgfsJob

All Known Implementing Classes:

GridGgfsInputStreamJobAdapter, GridGgfsJobAdapter

public interface GridGgfsJob

Defines executable unit for GridGgfsTask. Before this job is executed, it is assigned one of the ranges provided by the GridGgfsRecordResolver passed to one of the GridGgfs.execute(...) methods.

execute(GridGgfs, GridGgfsFileRange, GridGgfsInputStream) method is given GridGgfsFileRange this job is expected to operate on, and already opened GridGgfsInputStream for the file this range belongs to.

Note that provided input stream has position already adjusted to range start. However, it will not automatically stop on range end. This is done to provide capability in some cases to look beyond the range end or seek position before the reange start.

In majority of the cases, when you want to process only provided range, you should explicitly control amount of returned data and stop at range end. You can also use GridGgfsInputStreamJobAdapter, which operates on GridGgfsRangeInputStream bounded to range start and end, or manually wrap provided input stream with GridGgfsRangeInputStream.

You can inject any resources in concrete implementation, just as with regular GridComputeJob implementations.

Method Summary

Methods

Modifier and Type	Method and Description
void	cancel() This method is called when system detects that completion of this job can no longer alter the overall outcome (for example, when parent task has already reduced the results).
Object	execute(GridGgfs ggfs, GridGgfsFileRange range, GridGgfsInputStream in) Executes this job.

Method Detail

execute

Object execute(GridGgfs ggfs,
             GridGgfsFileRange range,
             GridGgfsInputStream in)
               throws GridException,
                      IOException

Executes this job.

Parameters:

ggfs - GGFS instance.

range - File range aligned to record boundaries.

in - Input stream for split file. This input stream is not aligned to range and points to file start by default.

Returns:

Execution result.

Throws:

GridException - If execution failed.

IOException - If file system operation resulted in IO exception.

cancel

void cancel()

This method is called when system detects that completion of this job can no longer alter the overall outcome (for example, when parent task has already reduced the results). Job is also cancelled when GridFuture.cancel() is called.

Note that job cancellation is only a hint, and just like with Thread.interrupt() method, it is really up to the actual job instance to gracefully finish execution and exit.