Moving this question to the dev list. Vladimir, any reason why we even allow peerClassLoadingEnabled set to true in IGFS/HadoopMR? I think the proper approach is to either ignore this setting or throw exception on startup, no?
D. On Wed, Dec 9, 2015 at 2:51 AM, Vladimir Ozerov <voze...@gridgain.com> wrote: > Mike, > > Thanks for attaching the project. It helped me got to the bottom of the > problem. Please remove the following line from example-igfs.xml file: > > <property name="peerClassLoadingEnabled" value="true"/> > > After that you should be able to use Hadoop Accelerator. This is a kind of > subtle usability issue. We will address it in the nearest feature ensuring > that meaningful error/warning is printed in such situations. > Please let me know if it resolves your problem. > > Vladimir. > > > On Wed, Dec 9, 2015 at 3:10 AM, mikep <michael.andre.pea...@me.com> wrote: > >> Hi Vladimir >> >> Heres a recreated example as standalone mini example with the same failing >> example in case it helps, recreating the same issue. >> >> Using intellij, >> >> run the IgniteNode class that starts up the ignite node (using spring). >> >> then run the TestExample class, >> >> this execute a few HDFS commands successfully and sets up a file on >> IgniteFS >> using HDFS api. >> and then executes the same job as mentioned before, on the cluster which >> suffers the issue. >> >> >> exampleignite.zip >> < >> http://apache-ignite-users.70518.x6.nabble.com/file/n2180/exampleignite.zip >> > >> >> just incase, java version is jdk1.8.0_25.jdk >> >> Cheers >> Mike >> >> >> >> -- >> View this message in context: >> http://apache-ignite-users.70518.x6.nabble.com/NPE-issue-with-trying-to-submit-Hadoop-MapReduce-tp2146p2180.html >> Sent from the Apache Ignite Users mailing list archive at Nabble.com. >> > >
