Hi All I'm trying to create a tiny 2-node cluster (both on linux FC7) with Hadoop 0.19.0 - previously, I was able to install and run hadoop on a single node. Now I'm trying it on 2 nodes - my idea was to put the name node and the job tracker on separate nodes, and initially use these two as the data nodes. So basically, the "master" and "slave" files - both have the names of these two nodes. When I start the dfs from the name node, it seems to work. But when I try to run the start-mapred.sh script, I get the following exception -
blueberry: Exception in thread "main" java.lang.NoClassDefFoundError: Could_not_reserve_enough_space_for_the_card_marking_array blueberry: Caused by: java.lang.ClassNotFoundException: Could_not_reserve_enough_space_for_the_card_marking_array blueberry: at java.net.URLClassLoader$1.run(URLClassLoader.java:200) blueberry: at java.security.AccessController.doPrivileged(Native Method) blueberry: at java.net.URLClassLoader.findClass(URLClassLoader.java:188) blueberry: at java.lang.ClassLoader.loadClass(ClassLoader.java:307) blueberry: at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301) blueberry: at java.lang.ClassLoader.loadClass(ClassLoader.java:252) blueberry: at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:320) blueberry: Could not find the main class: Could_not_reserve_enough_space_for_the_card_marking_array. Program will exit. Is it related to the heap space I allocated in the hadoop-env.sh? Or is it something else? Regards Arijit -- "And when the night is cloudy, There is still a light that shines on me, Shine on until tomorrow, let it be."