[WHIRR-687] Whirr 0.81 with Hadoop 0.23.x does not configure Hadoop cluster

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.8.1
• Component/s: service/hadoop

Environment

Amazon EC2 Ubuntu 12
Whirr 0.81
HBase 0.94.1
Hadoop 0.23.x

• Assignee: Unassigned
• Reporter: Deepak Vohra
• Created: 14/Dec/12 02:13
• Updated: 06/Feb/13 03:57

Description

The script that runs at Amazon EC2 instance startup with Whirr 0.81 does not configure Hadoop 0.23.x.

The default script that runs at instance startup on Amazon EC2 with Whirr configures Hadoop and HBase using directory structure from Hadoop 0.22. The directory structure in Hadoop 0.23.x is different than previous versions because of which an error gets generated.

At instance startup the following commands run to start the Hadoop cluster.
/usr/local/hadoop/bin/hadoop-daemon.sh start namenode
/usr/local/hadoop/bin/hadoop-daemon.sh start jobtracker
/usr/local/hadoop/bin/hadoop-daemon.sh start datanode
/usr/local/hadoop/bin/hadoop-daemon.sh start tasktracker

which generate error with Hadoop 0.23.x:
/usr/local/hadoop/bin/hadoop-daemon.sh: No such file or directory

For Hadoop 0.23.x the Whirr startup script should be modified for the different directory structure.