The yarn version of Giraph is quite finicky when it comes to detecting the giraph jar. When it can't find the jar on the new containers you get that exception.
After much experimentation I've had great success with the following (substitute paths and filenames where needed): HADOOP_CLASSPATH=/data/b.ajf/giraph hadoop jar /data/b.ajf/giraph/giraph.jar org.apache.giraph.GiraphRunner org.apache.giraph.examples.SimpleShortestPathsComputation -vip /user/b.ajf/ssp_input -vif org.apache.giraph.io.formats.JsonLongDoubleFloatDoubleVertexInputFormat -op /user/b.ajf/ssp_output -vof org.apache.giraph.io.formats.IdWithValueTextOutputFormat -yj giraph.jar The argument supplied to -yj must be only the name of the jar (no directories). The jar must be directly accessible from the current classpath or subdirectories (but it doesn't work if you give the full path of the jar, e.g, HADOOP_CLASSPATH=/data/b.ajf/giraph/giraph.jar). Also, if you use giraph-examples-with-dependencies you don't need the other one as the examples-with-dependencies already contain the core. Cheers, Alex On Thursday, February 20, 2014 05:42:45 PM Stefan Beskow wrote: > Thanks Roman for getting back to me so quickly. I checked with Cloudera and > they have a maven repository for CDH5 (version 2.2.0-cdh5.0.0-beta-2) as > shown below: > > <repository> > <id>cloudera</id> > <url>https://repository.cloudera.com/artifactory/cloudera-repos/</url> > </repository> > > I added that to the pom.xml file and ran a new build using the following > command: > > mvn -Phadoop_yarn -Dhadoop.version=2.2.0-cdh5.0.0-beta-2 clean package > -DskipTests > > The program now gets further, but throws the following exception: > > Container: container_1392713839733_0017_01_000002 on > el01cn04.unx.sas.com_8041 > =========================================================================== > ===== LogType: task-2-stderr.log > LogLength: 637 > Log Contents: > Exception in thread "main" java.lang.NoClassDefFoundError: > org/apache/giraph/yarn/GiraphYarnTask Caused by: > java.lang.ClassNotFoundException: org.apache.giraph.yarn.GiraphYarnTask at > java.net.URLClassLoader$1.run(URLClassLoader.java:202) > at java.security.AccessController.doPrivileged(Native Method) > at java.net.URLClassLoader.findClass(URLClassLoader.java:190) > at java.lang.ClassLoader.loadClass(ClassLoader.java:306) > at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301) > at java.lang.ClassLoader.loadClass(ClassLoader.java:247) > Could not find the main class: org.apache.giraph.yarn.GiraphYarnTask. > Program will exit. > > Do you have any additional suggestions for what I should try? > > Thanks for your help. > Stefan > > -----Original Message----- > From: shaposh...@gmail.com [mailto:shaposh...@gmail.com] On Behalf Of Roman > Shaposhnik Sent: Wednesday, February 19, 2014 4:43 PM > To: user@giraph.apache.org > Subject: Re: Problem running SimpleShortestPathsComputation on CDH5 with > Hadoop 2.2.0 > On Wed, Feb 19, 2014 at 1:12 PM, Stefan Beskow <stefan.bes...@sas.com> wrote: > > Hi. > > > > I checked out and built Giraph for Cloudera CDH5 with Hadoop 2.2.0 > > using the > > following: > > > > git clone git://git.apache.org/giraph.git snapshot_from_git cd > > snapshot_from_git mvn -Phadoop_yarn -Dhadoop.version=2.2.0 clean > > package -DskipTests > > It may help to build against CDH5 directly by: > * manually adding repository.cloudera.com to the set of repos > * specifying -Dhadoop.version=2.2.0-cdh5.0.0-beta-2 > > > When I run the sample application > > org.apache.giraph.examples.SimpleShortestPathsComputation I get the > > > following exception: > You need to provide way more logs from the YARN side for us to make sense of > it. > > Thanks, > Roman.
