/**
* splits the input files into tasks handled by a single node we have to read the input files to
* do this based on a number of items in a sequence
*/
@Override
public List<InputSplit> getSplits(JobContext job) throws IOException {
long filesPerTask = DistBlockFixer.filesPerTask(job.getConfiguration());
Path[] inPaths = getInputPaths(job);
List<InputSplit> splits = new ArrayList<InputSplit>();
long fileCounter = 0;
for (Path inPath : inPaths) {
FileSystem fs = inPath.getFileSystem(job.getConfiguration());
if (!fs.getFileStatus(inPath).isDir()) {
throw new IOException(inPath.toString() + " is not a directory");
}
FileStatus[] inFiles = fs.listStatus(inPath);
for (FileStatus inFileStatus : inFiles) {
Path inFile = inFileStatus.getPath();
if (!inFileStatus.isDir()
&& (inFile.getName().equals(job.getJobName() + IN_FILE_SUFFIX))) {
fileCounter++;
SequenceFile.Reader inFileReader =
new SequenceFile.Reader(fs, inFile, job.getConfiguration());
long startPos = inFileReader.getPosition();
long counter = 0;
// create an input split every filesPerTask items in the sequence
LongWritable key = new LongWritable();
Text value = new Text();
try {
while (inFileReader.next(key, value)) {
if (counter % filesPerTask == filesPerTask - 1L) {
splits.add(
new FileSplit(inFile, startPos, inFileReader.getPosition() - startPos, null));
startPos = inFileReader.getPosition();
}
counter++;
}
// create input split for remaining items if necessary
// this includes the case where no splits were created by the loop
if (startPos != inFileReader.getPosition()) {
splits.add(
new FileSplit(inFile, startPos, inFileReader.getPosition() - startPos, null));
}
} finally {
inFileReader.close();
}
}
}
}
LOG.info("created " + splits.size() + " input splits from " + fileCounter + " files");
return splits;
}
public DistBlockFixer(Configuration conf) {
super(conf);
filesPerTask = DistBlockFixer.filesPerTask(getConf());
maxPendingJobs = DistBlockFixer.maxPendingJobs(getConf());
maxFixTimeForFile = DistBlockFixer.maxFixTimeForFile(getConf());
}
