Connecting to Hive with Dremio 2.1

1

I am trying to connect to AWS EMR Single node instance with Hadoop/Hive running.

Connection works and I am even able to see hive tables. Strange thing though When I click on any tables to say query them it just keeps on running. The table nevertheless is very small, hardly a couple of rows.

Strangely there is nothing in the logs no errors nothing.

Any ideas as to why that might be the case.

2

Hi @drajput

You can try a few things,

  1. Look at the job profile on what the query is doing, if profile is unavailable then it is most probably doing a metadata refresh

Query Profile

  1. Try the same query with a LIMIT 1 or LIMIT 10

  2. If you can login to the Hive meta store using hive shell or beeline from the Dremio coordinator and see if you can select the table, that would be a good validation point

Thanks
@balaji.ramaswamy

3

Hi,

Well this is strange. See logs below from Dremio and Beeline.

  1. Beeline works fine.

57%20PM
Screen Shot 2018-09-21 at 4.20.57 PM.png2880×1082 761 KB

  1. Dremio is trying reconnection at <hadoop_ip>:8020.

55%20PM
Screen Shot 2018-09-21 at 4.26.55 PM.png2880×546 272 KB

53%20PM
Screen Shot 2018-09-21 at 4.35.53 PM.png1664×868 52.7 KB

4

Dremio is trying to read data that is stored on hdfs, so I assume that in your case it is connecting to wrong ip/hostname, is it correct?
Try to specify correct namenode(fs.defaultFS) in Advanced Options - it is on the left hand side on your screenshot