Something is wrong with your repositories, or with your Internet connection if you use repos on the Internet. Try to install just one of those packages directly from the command line, apt-get will show you more detailed error than Ambari, and then try to fix whatever is needed until apt-get install works, for example $ apt-get install hadoop-2-6-0-3-8-client Prepend "sudo" if needed. Also try to install a package unrelated to Hadoop like "tree": "sudo apt-get install tree". After both commands work, you can go back to Ambari and try to re-install the cluster. ... View more

There are no guarantees about region placement on region servers, even if all you RSs are running like before the "truncate". If some of them are not available HBase master will place regions on available ones. ... View more

What was your query? ... View more

You comment helped us too! However, on nodes running only ambari-agent the symlink should point to /usr/lib/ambari-agent/lib/resource_management ... View more

Yeah, that will be a lot of work, though Ambari maybe provides some automatism to create required paths based on service name, like /etc/hbase2/conf, /var/run/hbase2, /var/log/hbase2 etc. But still it doesn't sound like the best way to scale out services running on identical binaries (/usr/hdp/current/hbase2-client). Or maybe additional config files can be provided by /etc/hbase/conf2 pointing to /etc/hbase/HDP-VERSION/1? And by the way, we have this system of reputation points, upvoting or accepting helpfule replies. Can you please consider to use it on this post of mine? Tnx. ... View more

Not sure about Spark, but IMO you can do that when you configure HDFS, put SSD nodes in another Ambari config group and set the space not to be used by HDFS. ... View more

Okay, now I understand what do you mean by "cashing". Yes, you can remove RAID-1 on SSD's, then you can experiment with One_SSD and All_SSD policies, either way there are multiple replicas, so no need for RAID. And by the way there is no storage policy for NN, if possible it will be good to move 2x400G SSD from NN to worker nodes. ... View more

Change the namenode port in your job.properties, it should be: namenode=hdfs://sandbox.hortonworks.com:8020 You may have other errors too. By the way, there are many simple but good examples of Oozie actions in /usr/hdp/current/oozie-client/doc/examples on every Oozie client node. Copy that to your home directory and customize/test actions for your applications. Then copy "examples" to hdfs and test your actions. No need to write actions all by yourself. ... View more