[EDIT: Solved - but I am leaving the question in case anyone else hits the strange error message] I am upgrading a HDP cluster from 2.2 to 2.3 using a manual upgrade process. I have reached the end and perform the final step but ambari tells me that two of my machines are not upgraded. I can't see what the problem is, nor can I fix it. Any ideas? [root@hdp01 hdp]# ambari-server set-current --cluster-name=owalhdp --version-display-name=HDP-2.3.0.0 Using python /usr/bin/python2.6 Setting current version... Enter Ambari Admin login: admin Enter Ambari Admin password: ERROR: Exiting with exit code 1. REASON: Error during setting current version. Http status code - 500. { "status" : 500, "message" : "org.apache.ambari.server.controller.spi.SystemException: Finalization failed. More details: \nSTDOUT: Begin finalizing the upgrade of cluster owalhdp to version 2.3.0.0-2557\nThe following 2 host(s) have not been upgraded to version 2.3.0.0-2557. Please install and upgrade the Stack Version on those hosts and try again.\nHosts: hdp01.FQDN, hdp04.FQDN\n\nSTDERR: The following 2 host(s) have not been upgraded to version 2.3.0.0-2557. Please install and upgrade the Stack Version on those hosts and try again.\nHosts: hdp01.FQDN, hdp04.FQDN\n" } Now nothing in my manual process really tells me how to install and upgrade the Stack Version on individual hosts so I am at a loss for what to do. ... View more

I am thinking about setting up a Logstash infrastructure to monitor my system. (It happens to be Hortonworks HDP Hadoop cluster, but assume it isn't). So I have various things which generate logs and I want to transfer these logs outside my system to a new system - such as ElasticSearch inside Logstash. And I want to do this securely. I don't really want Flume for this as there are better tools. Now I might use Logstash forwarders - which most recently seems to be a new system called "Beats" - in particular FileBeat. However I would prefer to use Apache NiFi because of its security reputation. I would like to use HDF as I am a Hortonworks Partner and we are already using HDP. Can anyone say: "Yes this makes sense", "Yes, I have done it", "You need to read URL blah blah blah"? Or have I got the wrong end of the stick? PS I know that Ambari Metrics moves operational logs from the Hadoop cluster into the HDFS system - this is separate from that. ... View more

I've been trying to figure out what is the best/most appropriate Spark Notebook (web based IDE) for use on an HDP cluster. Is there one available through Hortonworks packages? I have found a number of training lessons from Hortonworks which talk about installing iPython Notebook (which can work for python+spark as well as scala+spark). However they seem to assume you are on a sandbox and are compiling and installing the software they demand, not checking what is already there and not using packages. I have also seen mentions of Zeppelin - an Apache incubating project. This is good too - but I have the same problem - I have to build it myself and it doesn't for me. There are also third party Notebooks such as the Cloud based one from Data Bricks. Any recommendations? What has worked for you? ... View more