On Tuesday 14 September 2010 02:00 AM, Harold Lim wrote:
Is there an instruction on how to use rumen?
Yes, the documentation was recently added to trunk via MAPREDUCE-1918:
https://issues.apache.org/jira/browse/MAPREDUCE-1918
I tried the following command:
java -cp
hadoop-mapred-0.21.0.jar:hadoop-common-0.21.0.jar:hadoop-hdfs-0.21.0.jar:hadoop-mapred-tools-0.21.0.jar:commons-logging-1.1.1.jar:commons-cli-1.2.jar:log4j-1.2.15.jar:jackson-mapper-asl-1.4.2.jar:jackson-core-asl-1.4.2.jar
org.apache.hadoop.tools.rumen.TraceBuilder traceout topologyout
10/09/13 16:26:09 INFO security.Groups: Group mapping
impl=org.apache.hadoop.security.ShellBasedUnixGroupsMapping; cacheTimeout=300000
10/09/13 16:26:10 WARN rumen.TraceBuilder: No job found in traces:
How do I set the input jobtracker log file?
Depending on the setting for "fs.default.name", this might be searching
in your HDFS home folder. It's better to specify the complete URL (using
"file:" or "hdfs:").
For example, suppose you want to create "jh.json" from the Job history
files in the "jobhistory" folder (all within the current directory).
You can then invoke Rumen like this:
$HADOOP_HOME/bin/hadoop jar
$HADOOP_HOME/hadoop-mapred-tools-0.22.0-SNAPSHOT.jar \
org.apache.hadoop.tools.rumen.TraceBuilder file://$PWD/jh.json
file://$PWD/topo.out \
file://$PWD/jobhistory
HTH,
Ranjit
