Hi All I have just started to use Hadoop (version 0.17.2) on a FC7 host with 2GB RAM. I've installed Sun JAVA 1.5.0_16 on the machine and successfully unpacked hadoop. I've also modified the paths in the conf/hadoop-env.sh file. But when I try to start hadoop using the start-all script, I get a JVM initialization error for the TaskTracker only - the rest of them start alright. I've set the JVM option (in HADOOP_OPTS) as -Xmx1024m.
strawberry:/home/hadoop> start-all.sh starting namenode, logging to /usr/local/hadoop/0.17.2/bin/../logs/hadoop-hadoop-namenode-strawberry.o ut JAVA_HOME is set to /usr/local/jdk1.5.0_16 localhost: starting datanode, logging to /usr/local/hadoop/0.17.2/bin/../logs/hadoop-hadoop-datanode-strawberry.o ut localhost: JAVA_HOME is set to /usr/local/jdk1.5.0_16 localhost: starting secondarynamenode, logging to /usr/local/hadoop/0.17.2/bin/../logs/hadoop-hadoop-secondarynamenode-str awberry.out localhost: JAVA_HOME is set to /usr/local/jdk1.5.0_16 starting jobtracker, logging to /usr/local/hadoop/0.17.2/bin/../logs/hadoop-hadoop-jobtracker-strawberry .out JAVA_HOME is set to /usr/local/jdk1.5.0_16 localhost: starting tasktracker, logging to /usr/local/hadoop/0.17.2/bin/../logs/hadoop-hadoop-tasktracker-strawberr y.out localhost: JAVA_HOME is set to /usr/local/jdk1.5.0_16 localhost: Could not create the Java virtual machine. localhost: Error occurred during initialization of VM localhost: Could not reserve enough space for object heap localhost: Could not create the Java virtual machine. Does the TaskTracker need more memory? The problem is if I increase the heap size in HADOOP_OPTS, all of the other hadoop processes start throwing the same error. Can anyone point me to the right direction please? Thanx in advance Arijit Dr. Arijit Mukherjee Principal Member of Technical Staff, Level-II Connectiva Systems (I) Pvt. Ltd. J-2, Block GP, Sector V, Salt Lake Kolkata 700 091, India Phone: +91 (0)33 23577531/32 x 107 http://www.connectivasystems.com
