Skip to content

Commit

Permalink
Creates continuous BulkIngest (#63)
Browse files Browse the repository at this point in the history
* A new Bulk import MapReduce job run using "cingest bulk"
* Will create a million random keys in Rfiles in the same format as CI
* The keys are in a linked list with the value being the prev row
* Works with CI createTable and verify
  • Loading branch information
milleruntime committed Mar 7, 2019
1 parent 65c55d0 commit 8f6c24c
Show file tree
Hide file tree
Showing 5 changed files with 283 additions and 11 deletions.
4 changes: 4 additions & 0 deletions bin/cingest
Expand Up @@ -33,6 +33,7 @@ Available applications:
verify Verifies continous ingest test. Stop ingest before running.
moru Streses Accumulo by reading and writing to the ingest table.
Stop ingest before running.
bulk Create RFiles in a Map Reduce job and calls importDirectory if successful
EOF
}

Expand Down Expand Up @@ -71,6 +72,9 @@ case "$1" in
moru)
ci_main="${ci_package}.ContinuousMoru"
;;
bulk)
ci_main="${ci_package}.BulkIngest"
;;
*)
echo "Unknown application: $1"
print_usage
Expand Down
5 changes: 5 additions & 0 deletions src/main/java/org/apache/accumulo/testing/TestProps.java
Expand Up @@ -32,6 +32,7 @@ public class TestProps {
private static final String CI_BW = CI + "batch.walker.";
private static final String CI_SCANNER = CI + "scanner.";
private static final String CI_VERIFY = CI + "verify.";
private static final String CI_BULK = CI + "bulk.";

/** Common properties **/
// HDFS root path. Should match 'fs.defaultFS' property in Hadoop's core-site.xml
Expand Down Expand Up @@ -116,6 +117,10 @@ public class TestProps {
// Location in HDFS to store output
public static final String CI_VERIFY_OUTPUT_DIR = CI_VERIFY + "output.dir";

/** Bulk **/
// Bulk ingest Job instance uuid
public static final String CI_BULK_UUID = CI_BULK + "uuid";

public static Properties loadFromFile(String propsFilePath) {
try {
return loadFromStream(new FileInputStream(propsFilePath));
Expand Down
259 changes: 259 additions & 0 deletions src/main/java/org/apache/accumulo/testing/continuous/BulkIngest.java
@@ -0,0 +1,259 @@
package org.apache.accumulo.testing.continuous;

import static org.apache.accumulo.testing.TestProps.CI_BULK_UUID;
import static org.apache.accumulo.testing.continuous.ContinuousIngest.genCol;
import static org.apache.accumulo.testing.continuous.ContinuousIngest.genLong;

import java.io.BufferedOutputStream;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import java.io.PrintStream;
import java.util.ArrayList;
import java.util.Base64;
import java.util.Collection;
import java.util.List;
import java.util.Random;
import java.util.UUID;

import org.apache.accumulo.core.client.AccumuloClient;
import org.apache.accumulo.core.data.Key;
import org.apache.accumulo.core.data.Value;
import org.apache.accumulo.hadoop.mapreduce.AccumuloFileOutputFormat;
import org.apache.accumulo.hadoop.mapreduce.partition.KeyRangePartitioner;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapreduce.InputFormat;
import org.apache.hadoop.mapreduce.InputSplit;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.JobContext;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.RecordReader;
import org.apache.hadoop.mapreduce.TaskAttemptContext;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
* Bulk import a million random key value pairs. Same format as ContinuousIngest and can be
* verified by running ContinuousVerify.
*/
public class BulkIngest extends Configured implements Tool {
public static final int NUM_KEYS = 1_000_000;
public static final String BULK_CI_DIR = "ci-bulk";

public static final Logger log = LoggerFactory.getLogger(BulkIngest.class);

@Override
public int run(String[] args) throws Exception {
Job job = Job.getInstance(getConf());
job.setJarByClass(BulkIngest.class);
// very important to prevent guava conflicts
job.getConfiguration().set("mapreduce.job.classloader", "true");
FileSystem fs = FileSystem.get(job.getConfiguration());

final String JOB_DIR = BULK_CI_DIR + "/" + getCurrentJobNumber(fs);
final String RFILE_DIR = JOB_DIR + "/rfiles";
log.info("Creating new job at {}", JOB_DIR);

String ingestInstanceId = UUID.randomUUID().toString();
job.getConfiguration().set(CI_BULK_UUID, ingestInstanceId);

log.info(String.format("UUID %d %s", System.currentTimeMillis(), ingestInstanceId));

Path outputDir = new Path(RFILE_DIR);

job.setInputFormatClass(RandomInputFormat.class);

// map the generated random longs to key values
job.setMapperClass(RandomMapper.class);
job.setMapOutputKeyClass(Key.class);
job.setMapOutputValueClass(Value.class);

// output RFiles for the import
job.setOutputFormatClass(AccumuloFileOutputFormat.class);
AccumuloFileOutputFormat.configure().outputPath(outputDir).store(job);

try (ContinuousEnv env = new ContinuousEnv(args)) {
String tableName = env.getAccumuloTableName();

// create splits file for KeyRangePartitioner
String splitsFile = JOB_DIR + "/splits.txt";
try (AccumuloClient client = env.getAccumuloClient()) {

// make sure splits file is closed before continuing
try (PrintStream out = new PrintStream(new BufferedOutputStream(fs.create(new Path(
splitsFile))))) {
Collection<Text> splits = client.tableOperations().listSplits(tableName, 100);
for (Text split : splits) {
out.println(Base64.getEncoder().encodeToString(split.copyBytes()));
}
job.setNumReduceTasks(splits.size() + 1);
}

job.setPartitionerClass(KeyRangePartitioner.class);
KeyRangePartitioner.setSplitFile(job, splitsFile);

job.waitForCompletion(true);
boolean success = job.isSuccessful();

// bulk import completed files
if (success) {
log.info("Sort and create job successful. Bulk importing {} to {}", RFILE_DIR, tableName);
client.tableOperations().importDirectory(RFILE_DIR).to(tableName).load();
} else {
log.error("Job failed, not calling bulk import");
}
return success ? 0 : 1;
}
}
}

private int getCurrentJobNumber(FileSystem fs) throws Exception {
Path jobPath = new Path(BULK_CI_DIR);
FileStatus jobDir = fs.getFileStatus(jobPath);
if (jobDir.isDirectory()) {
FileStatus[] jobs = fs.listStatus(jobPath);
return jobs.length;
} else {
log.info("{} directory doesn't exist yet, first job running will create it.", BULK_CI_DIR);
return 0;
}
}

public static void main(String[] args) throws Exception {
int ret = ToolRunner.run(new BulkIngest(), args);
System.exit(ret);
}

/**
* Mapper that takes the longs from RandomInputFormat and output Key Value pairs
*/
public static class RandomMapper extends Mapper<LongWritable,LongWritable,Key,Value> {

private String uuid;
private Text currentRow;
private Text currentValue;
private Text emptyCfCq;

@Override
protected void setup(Context context) {
uuid = context.getConfiguration().get(CI_BULK_UUID);
currentRow = new Text();
currentValue = new Text();
emptyCfCq = new Text(genCol(0));
}

@Override
protected void map(LongWritable key, LongWritable value, Context context) throws IOException,
InterruptedException {
currentRow.set(ContinuousIngest.genRow(key.get()));

// hack since we can't pass null - don't set first val (prevRow), we want it to be null
long longVal = value.get();
if (longVal != 1L) {
currentValue.set(ContinuousIngest.genRow(longVal));
}

Key outputKey = new Key(currentRow, emptyCfCq, emptyCfCq);
Value outputValue = ContinuousIngest.createValue(uuid.getBytes(), 0,
currentValue.copyBytes(), null);

context.write(outputKey, outputValue);
}
}

/**
* Generates a million LongWritable keys. The LongWritable value points to the previous key.
* The first key value pair has a value of 1L. This is translated to null in RandomMapper
*/
public static class RandomInputFormat extends InputFormat {

public static class RandomSplit extends InputSplit implements Writable {
@Override
public void write(DataOutput dataOutput) {}

@Override
public void readFields(DataInput dataInput) {}

@Override
public long getLength() {
return 0;
}

@Override
public String[] getLocations() {
return new String[0];
}
}

@Override
public List<InputSplit> getSplits(JobContext jobContext) {
List<InputSplit> splits = new ArrayList<>();
for (int i = 0; i < 10; i++) {
splits.add(new RandomSplit());
}
return splits;
}

@Override
public RecordReader createRecordReader(InputSplit inputSplit,
TaskAttemptContext taskAttemptContext) {
return new RecordReader() {
int number;
int currentNumber;
LongWritable currentKey;
LongWritable prevRow;
private Random random;

@Override
public void initialize(InputSplit inputSplit, TaskAttemptContext job) {
// number = Integer.parseInt(job.getConfiguration().get(NUM_KEYS));

This comment has been minimized.

Copy link
@mikewalch

mikewalch Mar 7, 2019

Member

should this code be commented out?

This comment has been minimized.

Copy link
@milleruntime

milleruntime Mar 7, 2019

Author Contributor

Yeah i left that in case we decide to make it configurable.

number = NUM_KEYS;
currentKey = new LongWritable(1);
prevRow = new LongWritable(1);
random = new Random();
currentNumber = 0;
}

@Override
public boolean nextKeyValue() {
if (currentNumber < number) {
prevRow.set(currentKey.get());
currentKey.set(genLong(0, Long.MAX_VALUE, random));
currentNumber++;
return true;
} else {
return false;
}
}

@Override
public LongWritable getCurrentKey() {
return currentKey;
}

@Override
public LongWritable getCurrentValue() {
return prevRow;
}

@Override
public float getProgress() {
return currentNumber * 1.0f / number;
}

@Override
public void close() throws IOException {

This comment has been minimized.

Copy link
@mikewalch

mikewalch Mar 7, 2019

Member

Is an empty close method expected here?

This comment has been minimized.

Copy link
@milleruntime

milleruntime Mar 7, 2019

Author Contributor

Yes this method is abstract but there is nothing to do.

}
};
}
}
}
Expand Up @@ -9,11 +9,11 @@
import org.apache.accumulo.testing.TestEnv;
import org.apache.accumulo.testing.TestProps;

class ContinuousEnv extends TestEnv {
public class ContinuousEnv extends TestEnv {

private List<Authorizations> authList;

ContinuousEnv(String[] args) {
public ContinuousEnv(String[] args) {
super(args);
}

Expand Down Expand Up @@ -43,23 +43,23 @@ Authorizations getRandomAuthorizations() {
return getAuthList().get(r.nextInt(getAuthList().size()));
}

long getRowMin() {
public long getRowMin() {
return Long.parseLong(testProps.getProperty(TestProps.CI_INGEST_ROW_MIN));
}

long getRowMax() {
public long getRowMax() {
return Long.parseLong(testProps.getProperty(TestProps.CI_INGEST_ROW_MAX));
}

int getMaxColF() {
public int getMaxColF() {
return Integer.parseInt(testProps.getProperty(TestProps.CI_INGEST_MAX_CF));
}

int getMaxColQ() {
public int getMaxColQ() {
return Integer.parseInt(testProps.getProperty(TestProps.CI_INGEST_MAX_CQ));
}

String getAccumuloTableName() {
public String getAccumuloTableName() {
return testProps.getProperty(TestProps.CI_COMMON_ACCUMULO_TABLE);
}
}
Expand Up @@ -249,8 +249,8 @@ public static Mutation genMutation(long rowLong, int cfInt, int cqInt, ColumnVis

byte[] rowString = genRow(rowLong);

byte[] cfString = FastFormat.toZeroPaddedString(cfInt, 4, 16, EMPTY_BYTES);
byte[] cqString = FastFormat.toZeroPaddedString(cqInt, 4, 16, EMPTY_BYTES);
byte[] cfString = genCol(cfInt);
byte[] cqString = genCol(cqInt);

if (checksum) {
cksum = new CRC32();
Expand All @@ -267,6 +267,10 @@ public static Mutation genMutation(long rowLong, int cfInt, int cqInt, ColumnVis
return m;
}

public static byte[] genCol(int cfInt) {
return FastFormat.toZeroPaddedString(cfInt, 4, 16, EMPTY_BYTES);
}

public static long genLong(long min, long max, Random r) {
return ((r.nextLong() & 0x7fffffffffffffffL) % (max - min)) + min;
}
Expand All @@ -275,11 +279,11 @@ static byte[] genRow(long min, long max, Random r) {
return genRow(genLong(min, max, r));
}

static byte[] genRow(long rowLong) {
public static byte[] genRow(long rowLong) {
return FastFormat.toZeroPaddedString(rowLong, 16, 16, EMPTY_BYTES);
}

private static Value createValue(byte[] ingestInstanceId, long count, byte[] prevRow,
public static Value createValue(byte[] ingestInstanceId, long count, byte[] prevRow,
Checksum cksum) {
int dataLen = ingestInstanceId.length + 16 + (prevRow == null ? 0 : prevRow.length) + 3;
if (cksum != null)
Expand Down

0 comments on commit 8f6c24c

Please sign in to comment.