org.apache.hadoop.hbase.mapreduce

Class TableSnapshotInputFormat

java.lang.Object

org.apache.hadoop.mapreduce.InputFormat<ImmutableBytesWritable,Result>

org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat

Direct Known Subclasses:

MultiTableSnapshotInputFormat

@InterfaceAudience.Public
public class TableSnapshotInputFormat
extends org.apache.hadoop.mapreduce.InputFormat<ImmutableBytesWritable,Result>

TableSnapshotInputFormat allows a MapReduce job to run over a table snapshot. The job bypasses HBase servers, and directly accesses the underlying files (hfile, recovered edits, wals, etc) directly to provide maximum performance. The snapshot is not required to be restored to the live cluster or cloned. This also allows to run the mapreduce job from an online or offline hbase cluster. The snapshot files can be exported by using the ExportSnapshot tool, to a pure-hdfs cluster, and this InputFormat can be used to run the mapreduce job directly over the snapshot files. The snapshot should not be deleted while there are jobs reading from snapshot files.

Usage is similar to TableInputFormat, and TableMapReduceUtil.initTableSnapshotMapperJob(String, Scan, Class, Class, Class, Job, boolean, Path) can be used to configure the job.

 Job job = new Job(conf);
 Scan scan = new Scan();
 TableMapReduceUtil.initTableSnapshotMapperJob(snapshotName,
      scan, MyTableMapper.class, MyMapKeyOutput.class,
      MyMapOutputValueWritable.class, job, true);
 
 

Internally, this input format restores the snapshot into the given tmp directory. By default, and similar to TableInputFormat an InputSplit is created per region, but optionally you can run N mapper tasks per every region, in which case the region key range will be split to N sub-ranges and an InputSplit will be created per sub-range. The region is opened for reading from each RecordReader. An internal RegionScanner is used to execute the CellScanner obtained from the user.

HBase owns all the data and snapshot files on the filesystem. Only the 'hbase' user can read from snapshot files and data files. To read from snapshot files directly from the file system, the user who is running the MR job must have sufficient permissions to access snapshot and reference files. This means that to run mapreduce over snapshot files, the MR job has to be run as the HBase user or the user must have group or other privileges in the filesystem (See HBASE-8369). Note that, given other users access to read from snapshot/data files will completely circumvent the access control enforced by HBase.

See Also:

TableSnapshotScanner

Constructor Summary

Constructors

Constructor and Description
TableSnapshotInputFormat()

Method Summary

Modifier and Type	Method and Description
org.apache.hadoop.mapreduce.RecordReader<ImmutableBytesWritable,Result>	createRecordReader(org.apache.hadoop.mapreduce.InputSplit split, org.apache.hadoop.mapreduce.TaskAttemptContext context)
List<org.apache.hadoop.mapreduce.InputSplit>	getSplits(org.apache.hadoop.mapreduce.JobContext job)
static void	setInput(org.apache.hadoop.mapreduce.Job job, String snapshotName, org.apache.hadoop.fs.Path restoreDir) Configures the job to use TableSnapshotInputFormat to read from a snapshot.
static void	setInput(org.apache.hadoop.mapreduce.Job job, String snapshotName, org.apache.hadoop.fs.Path restoreDir, org.apache.hadoop.hbase.util.RegionSplitter.SplitAlgorithm splitAlgo, int numSplitsPerRegion) Configures the job to use TableSnapshotInputFormat to read from a snapshot.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TableSnapshotInputFormat

public TableSnapshotInputFormat()

Method Detail

createRecordReader

public org.apache.hadoop.mapreduce.RecordReader<ImmutableBytesWritable,Result> createRecordReader(org.apache.hadoop.mapreduce.InputSplit split,
                                                                                                  org.apache.hadoop.mapreduce.TaskAttemptContext context)
                                                                                           throws IOException

Specified by:

createRecordReader in class org.apache.hadoop.mapreduce.InputFormat<ImmutableBytesWritable,Result>

Throws:

IOException

getSplits

public List<org.apache.hadoop.mapreduce.InputSplit> getSplits(org.apache.hadoop.mapreduce.JobContext job)
                                                       throws IOException,
                                                              InterruptedException

Specified by:

getSplits in class org.apache.hadoop.mapreduce.InputFormat<ImmutableBytesWritable,Result>

Throws:

IOException

InterruptedException

setInput

public static void setInput(org.apache.hadoop.mapreduce.Job job,
                            String snapshotName,
                            org.apache.hadoop.fs.Path restoreDir)
                     throws IOException

Configures the job to use TableSnapshotInputFormat to read from a snapshot.

Parameters:

job - the job to configure

snapshotName - the name of the snapshot to read from

restoreDir - a temporary directory to restore the snapshot into. Current user should have write permissions to this directory, and this should not be a subdirectory of rootdir. After the job is finished, restoreDir can be deleted.

Throws:

IOException - if an error occurs

setInput

public static void setInput(org.apache.hadoop.mapreduce.Job job,
                            String snapshotName,
                            org.apache.hadoop.fs.Path restoreDir,
                            org.apache.hadoop.hbase.util.RegionSplitter.SplitAlgorithm splitAlgo,
                            int numSplitsPerRegion)
                     throws IOException

Configures the job to use TableSnapshotInputFormat to read from a snapshot.

Parameters:

job - the job to configure

snapshotName - the name of the snapshot to read from

restoreDir - a temporary directory to restore the snapshot into. Current user should have write permissions to this directory, and this should not be a subdirectory of rootdir. After the job is finished, restoreDir can be deleted.

splitAlgo - split algorithm to generate splits from region

numSplitsPerRegion - how many input splits to generate per one region

Throws:

IOException - if an error occurs