amplab / graphx

Former GraphX development repository. GraphX has been merged into Apache Spark; please submit pull requests there.
https://github.com/apache/spark
Apache License 2.0
360 stars 103 forks source link

There are some issues in article "Launch a benchmarking cluster" #139

Open qianlong-zhang opened 10 years ago

qianlong-zhang commented 10 years ago
I want to run pagerank on Graphx, following the instructions(https://github.com/amplab/graphx/wiki/Launch-a-benchmarking-cluster), I encountered some problems.
First, the running command(~/graphx/run-example org.apache.spark.graph.Analytics spark://$MASTERS:7077 pagerank hdfs://$MASTERS:9000/soc-LiveJournal1.txt --numIter=20 --numEPart=128) is wrong, I changed it to "./bin/run-example org.apache.spark.graphx.lib.Analytics spark://XXX:7077 pagerank hdfs://XXX:8020/soc-LiveJournal1.txt". The parameter "--numIter" can not be found in running pagerank, I read the source code and find it is used in cc benchmark.

when I run the command above, spark throw warnings and errors: 14/04/21 10:24:59 WARN scheduler.TaskSchedulerImpl: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient memory …… Exception in thread "main" org.apache.spark.SparkException: Job aborted due to stage failure: Spark cluster looks down

What I am sure is spark UI is ok, and other benchmarks can run rightly.

What I expect is, if you can give some hint in deploying Graphx on standlone cluster? Thanks very very much, cause I have been blocked on this for 1 week. :)