Spark through the BulkLoad quickly on the massive data into the Hbase

We introduced a quick way to import massive amounts of data into Hbase by introducing bulk data into Hbase [Hadoop articles] through BulkLoad . This article will show you how to use Scala to quickly import data into Springs Methods. Here are two ways: the first to use Put ordinary method to count down; the second use Bulk Load API. For more information on why you need to use Bulk Load This article is not introduced , see "BulkLoad quickly into the massive data into Hbase [Hadoop articles]" .

If you want to keep abreast of Spark , Hadoop or Hbase related articles, please pay attention to WeChat public account: iteblog_hadoop

Use org.apache.hadoop.hbase.client.Put to write data

Use org.apache.hadoop.hbase.client.Put to write data one by one to the Hbase, but it is org.apache.hadoop.hbase.client.Put than Bulk loading, just as a contrast.
import org.apache.spark._
import org.apache.spark.rdd.NewHadoopRDD
import org.apache.hadoop.hbase.{HBaseConfiguration, HTableDescriptor}
import org.apache.hadoop.hbase.client.HBaseAdmin
import org.apache.hadoop.hbase.mapreduce.TableInputFormat
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hbase.HColumnDescriptor
import org.apache.hadoop.hbase.util.Bytes
import org.apache.hadoop.hbase.client.Put;
import org.apache.hadoop.hbase.client.HTable;
  
val conf = HBaseConfiguration.create()
val tableName = "/iteblog"
conf.set(TableInputFormat.INPUT_TABLE, tableName)
  
val myTable = new HTable(conf, tableName);
var p = new Put();
p = new Put(new String("row999").getBytes());
p.add("cf".getBytes(), "column_name".getBytes(), new String("value999").getBytes());
myTable.put(p);
myTable.flushCommits();

Batch guide to Hbase

Batch guide data to the Hbase can be divided into two kinds: (1), generate Hfiles, and then batch guide data;
(2), direct import of data into the Hbase.

Bulk imports Hfiles into Hbase

Now we have to introduce how to batch data into the Hbase, mainly divided into two steps:
(1), Mr. Hfiles;
(2), use org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles to import Hfiles into Hbase in advance.
The implementation code is as follows:
import org.apache.spark._
import org.apache.spark.rdd.NewHadoopRDD
import org.apache.hadoop.hbase.{HBaseConfiguration, HTableDescriptor}
import org.apache.hadoop.hbase.client.HBaseAdmin
import org.apache.hadoop.hbase.mapreduce.TableInputFormat
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hbase.HColumnDescriptor
import org.apache.hadoop.hbase.util.Bytes
import org.apache.hadoop.hbase.client.Put;
import org.apache.hadoop.hbase.client.HTable;
import org.apache.hadoop.hbase.mapred.TableOutputFormat
import org.apache.hadoop.mapred.JobConf
import org.apache.hadoop.hbase.io.ImmutableBytesWritable
import org.apache.hadoop.mapreduce.Job
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat
import org.apache.hadoop.hbase.KeyValue
import org.apache.hadoop.hbase.mapreduce.HFileOutputFormat
import org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles
  
val conf = HBaseConfiguration.create()
val tableName = "iteblog"
val table = new HTable(conf, tableName)
  
conf.set(TableOutputFormat.OUTPUT_TABLE, tableName)
val job = Job.getInstance(conf)
job.setMapOutputKeyClass (classOf[ImmutableBytesWritable])
job.setMapOutputValueClass (classOf[KeyValue])
HFileOutputFormat.configureIncrementalLoad (job, table)
  
// Generate 10 sample data:
val num = sc.parallelize(1 to 10)
val rdd = num.map(x=>{
    val kv: KeyValue = new KeyValue(Bytes.toBytes(x), "cf".getBytes(), "c1".getBytes(), "value_xxx".getBytes() )
    (new ImmutableBytesWritable(Bytes.toBytes(x)), kv)
})
  
// Save Hfiles on HDFS
rdd.saveAsNewAPIHadoopFile("/tmp/iteblog", classOf[ImmutableBytesWritable], classOf[KeyValue], classOf[HFileOutputFormat], conf)
  
//Bulk load Hfiles to Hbase
val bulkLoader = new LoadIncrementalHFiles(conf)
bulkLoader.doBulkLoad(new Path("/tmp/iteblog"), table)
After running the above code, we can see that the itblog table in the Hbase has generated 10 data, as follows:
hbase(main):020:0> scan 'iteblog'
ROW                                                 COLUMN+CELL
 \x00\x00\x00\x01                                   column=cf:c1, timestamp=1425128075586, value=value_xxx
 \x00\x00\x00\x02                                   column=cf:c1, timestamp=1425128075586, value=value_xxx
 \x00\x00\x00\x03                                   column=cf:c1, timestamp=1425128075586, value=value_xxx
 \x00\x00\x00\x04                                   column=cf:c1, timestamp=1425128075586, value=value_xxx
 \x00\x00\x00\x05                                   column=cf:c1, timestamp=1425128075586, value=value_xxx
 \x00\x00\x00\x06                                   column=cf:c1, timestamp=1425128075675, value=value_xxx
 \x00\x00\x00\x07                                   column=cf:c1, timestamp=1425128075675, value=value_xxx
 \x00\x00\x00\x08                                   column=cf:c1, timestamp=1425128075675, value=value_xxx
 \x00\x00\x00\x09                                   column=cf:c1, timestamp=1425128075675, value=value_xxx
 \x00\x00\x00\x0A                                   column=cf:c1, timestamp=1425128075675, value=value_xxx

Direct Bulk Load data to Hbase

This method does not need to generate Hfiles on HDFS in advance, but directly to the bulk of the data into the Hbase. There are only minor differences compared to the above example, as follows:
will
rdd.saveAsNewAPIHadoopFile("/tmp/iteblog", classOf[ImmutableBytesWritable], classOf[KeyValue], classOf[HFileOutputFormat], conf) 
changed to:
rdd.saveAsNewAPIHadoopFile("/tmp/iteblog", classOf[ImmutableBytesWritable], classOf[KeyValue], classOf[HFileOutputFormat], job.getConfiguration())
The complete implementation is as follows:
import org.apache.spark._
import org.apache.spark.rdd.NewHadoopRDD
import org.apache.hadoop.hbase.{HBaseConfiguration, HTableDescriptor}
import org.apache.hadoop.hbase.client.HBaseAdmin
import org.apache.hadoop.hbase.mapreduce.TableInputFormat
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hbase.HColumnDescriptor
import org.apache.hadoop.hbase.util.Bytes
import org.apache.hadoop.hbase.client.Put;
import org.apache.hadoop.hbase.client.HTable;
import org.apache.hadoop.hbase.mapred.TableOutputFormat
import org.apache.hadoop.mapred.JobConf
import org.apache.hadoop.hbase.io.ImmutableBytesWritable
import org.apache.hadoop.mapreduce.Job
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat
import org.apache.hadoop.hbase.KeyValue
import org.apache.hadoop.hbase.mapreduce.HFileOutputFormat
import org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles
  
val conf = HBaseConfiguration.create()
val tableName = "iteblog"
val table = new HTable(conf, tableName)
  
conf.set(TableOutputFormat.OUTPUT_TABLE, tableName)
val job = Job.getInstance(conf)
job.setMapOutputKeyClass (classOf[ImmutableBytesWritable])
job.setMapOutputValueClass (classOf[KeyValue])
HFileOutputFormat.configureIncrementalLoad (job, table)
  
// Generate 10 sample data:
val num = sc.parallelize(1 to 10)
val rdd = num.map(x=>{
    val kv: KeyValue = new KeyValue(Bytes.toBytes(x), "cf".getBytes(), "c1".getBytes(), "value_xxx".getBytes() )
    (new ImmutableBytesWritable(Bytes.toBytes(x)), kv)
})
  
// Directly bulk load to Hbase/MapRDB tables.
rdd.saveAsNewAPIHadoopFile("/tmp/iteblog", classOf[ImmutableBytesWritable], classOf[KeyValue], classOf[HFileOutputFormat], job.getConfiguration())

other

In the above example, we used the saveAsNewAPIHadoopFile API to write data to HBase; in fact, we can also use the saveAsNewAPIHadoopDataset API to achieve the same goal, we only need to the saveAsNewAPIHadoopDataset code
rdd.saveAsNewAPIHadoopFile("/tmp/iteblog", classOf[ImmutableBytesWritable], classOf[KeyValue], classOf[HFileOutputFormat], job.getConfiguration())
changed to
job.getConfiguration.set("mapred.output.dir", "/tmp/iteblog")
rdd.saveAsNewAPIHadoopDataset(job.getConfiguration)
The rest is exactly the same as before.

Commentaires

Enregistrer un commentaire

Posts les plus consultés de ce blog

Controlling Parallelism in Spark by controlling the input partitions by controlling the input partitions

Image
It is useful to be able to control the degree of parallelism when using Spark. Spark provides a very convenient method to increase the degree of parallelism which should be adequate in practice. This blog entry of the "Inside Spark" series describes the knobs Spark uses to control the degree of parallelism. Controlling the number of partitions in the " local " mode Spark is designed ground up to enable unit testing. To control the degree of parallelism in the local mode simply utilize the function JavaRDD<String> org.apache.spark.api.java.JavaSparkContext.parallelize(List<String> list, int numSlices) The following snippet of code demonstrates this usage- public class SparkControlPartitionSizeLocally { public static void main ( String [] args ) { JavaSparkContext sc = new JavaSparkContext ( "local" , "localpartitionsizecontrol" ); String input = "four score and sev
Lire la suite

Spark performance optimization: shuffle tuning

Image
Article directory 1 shuffle tuning 1.1 Summary of tuning 1.2 Overview of ShuffleManager Development 1.3 HashShuffleManager operating principle 1.3.1 Unoptured HashShuffleManager 1.3.2 Optimized HashShuffleManager 1.4 SortShuffleManager operating principle 1.4.1 General operating mechanism 1.4.2 bypass running mechanism 1.5 shuffle related parameters tuning 1.5.1 spark.shuffle.file.buffer 1.5.2 spark.reducer.maxSizeInFlight 1.5.3 spark.shuffle.io.maxRetries 1.5.4 spark.shuffle.io.retryWait 1.5.5 spark.shuffle.memoryFraction 1.5.6 spark.shuffle.manager 1.5.7 spark.shuffle.sort.bypassMergeThreshold 1.5.8 spark.shuffle.consolidateFiles 2 write in the last words Shuffle Summary of tuning Most of the performance of Spark operations is mainly consumed in the shuffle link, because the link contains a large number of disk IO, serialization, network data transmission and othe
Lire la suite

Spark optimization

Image
One operation and maintenance 1. Master hang up, standby restart is also invalid Master defaults to 512M of memory, when the task in the cluster is particularly high, it will hang, because the master will read each task event log log to generate spark ui, the memory will naturally OOM, you can run the log See that the master of the start through the HA will naturally fail for this reason. solve Increase the Master's memory spark-env.sh , set in the master node spark-env.sh : export SPARK_DAEMON_MEMORY 10g # 根据你的实际情况 Reduce the job information stored in the Master memory spark.ui.retainedJobs 500 # 默认都是1000 spark.ui.retainedStages 500 Hang up or suspend Sometimes we will see the web node in the web ui disappear or in the dead state, the task of running the node will report a variety of lost worker errors, causing the same reasons and the above, worker memory to save a lot of ui The information leads to gc when the heartbeat is lost
Lire la suite