Hi All,
Saw en error while launching Hadoop cluster, details are
Whirr version: 0.7.0
Using Cloudera image
Bootstrapping a simple cluster for playing, observed this
error(Cluster runs fine)
Starting Hadoop jobtracker daemon (hadoop-jobtracker):
startingjobtracker, logging
to/var/log/hadoop/logs/hadoop-hadoop-jobtracker-ip-XXXXXX.out,
error=/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a
directory/etc/profile: line 58: /usr/java/jdk1.6.0_21: is a directory,
exitCode=0]
Couple of questions
1. How can I grant privileges to the login user so that I can
controlall the stuff running on cluster
2. Once the cluster is running, a display summary would be better like
Name Node IP Login string
Data Node IP Login String ....
It will make life easy for the user :)
3. Is it possible to restart the cluster instance manually
afterlogging in or the only way is to destroy and launch the cluster
again?
4. Is it possible to use whirr to configure a Hadoop cluster
withalready running instances on EC2, rather than firing new ones? The
Usecase is to reuse instances during off-peak hours, which would
anywaysbe running.
--
thanks
ashish
Blog: http://www.ashishpaliwal.com/blog
My Photo Galleries: http://www.pbase.com/ashishpaliwal
