I think this is a real problem we need to address soon - maybe even make a new release (0.7.2).
Can you try to change the install java function to make it work for you? On Feb 29, 2012 7:04 PM, "Sebastian Schoenherr" < [email protected]> wrote: > Hi Andrei, > thanks a lot for your reply. No, the hadoop service is not starting as > expected. I added the suggested line to the end of my property file but the > result is still the same. I'll keep on trying. > > Here is the log from /tmp/logs/stderr.log, it looks like my > hadoop-env.JAVA_HOME is not used. > + export JAVA_HOME=/usr/lib/jvm/java-1.6.0-openjdk > + JAVA_HOME=/usr/lib/jvm/java-1.6.0-openjdk > + echo 'export JAVA_HOME=/usr/lib/jvm/java-1.6.0-openjdk' > + echo 'export JAVA_HOME=/usr/lib/jvm/java-1.6.0-openjdk' > + alternatives --install /usr/bin/java java > /usr/lib/jvm/java-1.6.0-openjdk/bin/java 17000 > + alternatives --set java /usr/lib/jvm/java-1.6.0-openjdk/bin/java > + java -version > /tmp/setup-user.sh: line 247: java: command not found > + exit 1 > > Thanks, > Sebastian > > On 29.02.2012 17:05, Andrei Savu wrote: > > >> thanks for the WHIRR Update. I just started a cluster with an Ubuntu >> image and it works great. Unfortunately, I run into some problems when >> trying to set up a cluster with the basic Amazon Images (AMI 2011/09, 32bit >> and 64bit). I get an 'java command not found error' on the instances. >> > > But is Hadoop starting as expected? > > No, hadoop is not starting as expected > > > >> I think the problem is that the Java path is different to the ubuntu >> images. On 64Bit the correct path would be >> /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64 instead of >> /usr/lib/jvm/java-1.6.0-openjdk. >> > > We've done testing only on Ubuntu 10.04 LTS and I think you are right. > You can workaround this limitation by adding something like this to your > properties file: > > hadoop-env.JAVA_HOME= /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64 > > >> >> Maybe I'm missing something out, here is my property file: >> Thanks for your help, >> Sebastian >> >> >> whirr.cluster-name = 1330 >> whirr.instance-templates = 1 hadoop-jobtracker+hadoop-namenode,1 >> hadoop-datanode+hadoop-tasktracker >> > > This should be hadoop-namenode+hadoop-jobtracker *not* the other way > around. > > >> whirr.cluster-user = my-user >> whirr.provider = aws-ec2 >> whirr.image-id = us-east-1/ami-31814f58 >> whirr.login-user = ec2-user >> > > This option is not required. > > >> whirr.hardware-id = t1.micro >> > > I recommend you to use at least m1.small - t1.micro is less than ideal > in this case. > > >> whirr.private-key-file = ... >> whirr.public-key-file = ... >> whirr.identity = ... >> whirr.credential = ... >> whirr.hadoop.install-function = install_cdh_hadoop >> whirr.hadoop.configure-function = configure_cdh_hadoop >> >> >> >> On 29.02.2012 09:44, Andrei Savu wrote: >> >>> The Apache Whirr team is pleased to announce the release of Apache Whirr >>> 0.7.1. >>> >>> Whirr is a library and a command line tool that can be used to run >>> distributed >>> services in the cloud. It simplifies the deployment of distributed >>> systems on >>> cloud infrastructure, allowing you to launch and tear-down complex >>> cloud cluster >>> environments with a single command. >>> >>> Supported services currently include most of the components of the Apache >>> Hadoop stack, Apache Mahout, Chef, Puppet, Ganglia, elasticsearch, Apache >>> Cassandra, Voldemort and Hama. Services can be deployed to Amazon EC2 >>> and to Rackspace Cloud. >>> >>> The release is available here: >>> http://www.apache.org/dyn/closer.cgi/whirr/ >>> >>> The full change log is available here: >>> https://issues.apache.org/jira/browse/WHIRR/fixforversion/12319942 >>> >>> We welcome your help and feedback. For more information on how to >>> report problems, and to get involved, visit the project website at >>> http://whirr.apache.org/ >>> >>> The Apache Whirr Team >>> >> > >
