big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

MAPREDUCE-2398. MRBench: setting the baseDir parameter has no effect. Contributed by Michael Noll and Wilfred Spiegelenburg.

Browse Source
This commit is contained in:
Harsh J 2016-04-21 17:33:22 +05:30
parent 170c4fd4cd
commit 7da5847cf1
1 changed files with 12 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/test/java/org/apache/hadoop/mapred/MRBench.java
View File

@ -41,10 +41,13 @@
public class MRBench extends Configured implements Tool{
private static final Log LOG = LogFactory.getLog(MRBench.class);
private static final String DEFAULT_INPUT_SUB = "mr_input";
private static final String DEFAULT_OUTPUT_SUB = "mr_output";
private static Path BASE_DIR =
new Path(System.getProperty("test.build.data","/benchmarks/MRBench"));
private static Path INPUT_DIR = new Path(BASE_DIR, "mr_input");
private static Path OUTPUT_DIR = new Path(BASE_DIR, "mr_output");
private static Path INPUT_DIR = new Path(BASE_DIR, DEFAULT_INPUT_SUB);
private static Path OUTPUT_DIR = new Path(BASE_DIR, DEFAULT_OUTPUT_SUB);
public static enum Order {RANDOM, ASCENDING, DESCENDING};
@ -243,6 +246,8 @@ public int run(String[] args) throws Exception {
numRuns = Integer.parseInt(args[++i]);
} else if (args[i].equals("-baseDir")) {
BASE_DIR = new Path(args[++i]);
INPUT_DIR = new Path(BASE_DIR, DEFAULT_INPUT_SUB);
OUTPUT_DIR = new Path(BASE_DIR, DEFAULT_OUTPUT_SUB);
} else if (args[i].equals("-maps")) {
numMaps = Integer.parseInt(args[++i]);
} else if (args[i].equals("-reduces")) {
@ -283,14 +288,15 @@ public int run(String[] args) throws Exception {
Path inputFile = new Path(INPUT_DIR, "input_" + (new Random()).nextInt() + ".txt");
generateTextFile(fs, inputFile, inputLines, inputSortOrder);
// setup test output directory
fs.mkdirs(BASE_DIR);
ArrayList<Long> execTimes = new ArrayList<Long>();
try {
execTimes = runJobInSequence(jobConf, numRuns);
} finally {
// delete output -- should we really do this?
fs.delete(BASE_DIR, true);
// delete all generated data -- should we really do this?
// we don't know how much of the path was created for the run but this
// cleans up as much as we can
fs.delete(OUTPUT_DIR, true);
fs.delete(INPUT_DIR, true);
}
if (verbose) {