I don't see a reason for the first insert to be a text/uncompressed avro file. Using HCatalog, you can directly import from sqoop to hive table as ORC. That would save you a lot of space because of compression. Once the initial data import is in Hive as ORC, you can then still continue and transform this data as necessary. If the reason for writing as text is to access from Pig and MR, a HCatalog table also can be accessed from Pig/MR. ... View more

If you already have a running cluster, exporting blueprint from ambari, editing relevant entries and using that to create a DR cluster works. Another approach is to create your first cluster as well with a blueprint. This will ensure easier creation of second cluster. I don't know of any other custom tools that can create an almost identical cluster ... View more

Checkpointing is the process of merging editlogs with base fsimage. This will be stored in namenode metadata directories. Its not the same as editlog, since editlog has the changes that you make to HDFS. ... View more

When you click on OVA to import, you will see Guest OS Type? What do you see there and has this been changed. You should see Red Hat (64-bit) there. By default you don't change this and if you change this, import is not going to work. ... View more