Setup Hadoop Nodes

• setup node to pxe boot
• SSH to ics.lsdf.kit.edu (ask Ursula Epting)
• Go to directory /srv/tftp/pxelinux.cfg/lsdf/
• Rename link mv hph-c-x.disabled hph-c-x
• Go to directory /srv/tftp/pxelinux.cfg/
• Create link ln -s lsdf/hph-c-0XX YYYYYYYY where XX is the node number, YYYYYYYY is the IP address (sedecimal)
• Reboot client, monitor on console
• After installation finished, invoke /usr/local/sbin/cfe-enable-client.sh hph-c-0XX.lsdf.kit.edu
• Check /etc/cfengine/cfrun.hosts whether the node is included
• Invoke cfrun hph-c-0XX.lsdf.kit.edu -- -q
• Go to node, check if .ssh directory exists.
• Invoke cfagent --verbose --debug locally on client, if not
• Invoke on client yum --disableexcludes=all update
• /usr/local/sbin/postinst-packages-cleanup.sh
• yum install -y --disableexcludes=all hadoop-0.20 hadoop-0.20-datanode hadoop-0.20-tasktracker hadoop-0.20-fuse.x86_64 hadoop-0.20-libhdfs.x86_64 hadoop-0.20-native.x86_64 hadoop-0.20-pipes.x86_64 hadoop-0.20-sbin.x86_64 hadoop-hbase hadoop-hbase-regionserver hadoop-zookeeper
• /usr/local/sbin/postinst-network-bridge-setup.sh
• The cfengine configuration misses two essential steps: chkconfig --list jexec and ln -s /usr/java/jdk1.6.0_41 /usr/java/latest
• After a reboot (due to kernel update) You should check in this order /etc/init.d/hadoop-hbase-regionserver status with result HBase is running, /etc/init.d/hadoop-0.20-datanode status resulting datanode (PID 5359) wird ausgeführt... and /etc/init.d/hadoop-0.20-tasktracker status