hi Kamal,

All the hadoop services will be started when you start cloudera VM. if in case you see that there is connection errors to hdfs here is what you can do:

- login to cloudera manager in firefox, using username/password: cloudera/cloudera
- once you login the user will be taken to "services" page
- under "services" page check the status of "hdfs1", its the icon nex to it shows "Green" then you are all set, if you see "red" instead goto "actions" on the same page for "hdfs1" and click restart.

- once the service is restarted, fire the hdfs commands on the terminal and verify the output.

Thanks

As an FYI, notice the following in your log.

Directory /var/lib/hadoop-0.20/cache/hadoop/dfs/name is in an
inconsistent state: storage directory does not exist or is not
accessible.

Your version of CDH uses hadoop 0.2 The instructor told us to explicitly use 1.x of hadoop. I did a little research and found that there are NO CDH packages that use hadoop 1.x. Only 0.20 or 2.x. Which brings out the question as to whether we should be using a CDH package for this course. It may be possible to get one of the packages and install hadoop 1.x on it seeing that there may be value in using the other tools that are installed like pig and hive. However, if that is the road you take, I would suggest looking at CDH 4 which is a 2.x of hadoop since it is more likely that the other tools would be backward compatible to hadoop 1.x whereas I don't know if those tools meant to run on 0.20 would run on a 2.x version of hadoop.

I think I am going to play it safe and install from scratch. I'll learn more from it anyway.

Thanks

I also got a JVM error on my CDH3 installed.

Where can I find the detailed step by step instructions for installing from scratch on a windows system?

Thanks

How to access clodera manager? Can you tell me the URL?

Looks like you need to run namenode -format first to create storage directory