Which tarballs are missing? They are uploaded to HDFS when certain services start, e.g., Hive Server, History Server, Tez Service Check, Spark Service Check. You can add logging to the method copy_to_hdfs in copy_tarball.py to ensure it is called. Even after installation, you can perform these ops to try uploading tarballs.
For VMs deployed via blueprints, I'm not quite sure if we don't run those steps to improve start-up time. Thanks, Alejandro From: Andrey Klochkov <[email protected]<mailto:[email protected]>> Reply-To: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Date: Tuesday, June 14, 2016 at 2:19 PM To: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Ambari not creating /hdp/apps/<hdp-version> when using a blueprint Hi, Should Ambari upload tarballs to /hdp/apps/<hdp-version> in HDFS when creating a cluster from a blueprint? Seems it doesn't do that in our case. I see that there's Ambaripreupload.py script in Ambari sources that does that but can't figure out how that script is supposed to be executed. No references in source code. Nothing in the logs. How can I troubleshoot that? Thanks! -- Andrey Klochkov
