Hive connection trouble

Hi!
I’m having trouble connecting Dremio Community to a remote Hive database. The error I get is:

java.net.ConnectException: Call From tqbigdprodldremiomr01/10.24.192.243 to isilonscprodhadoop.sari.local:8020 failed on connection exception: java.net.ConnectException: Connection refused; For more details see: http://wiki.apache.org/hadoop/ConnectionRefused

The Hive instance connects and I actually get the tables listed from the metastore. The problem comes with every query.
Every other connection: Greenplum, HDFS, ElasticSearch … works just fine.
Our HDFS is virtualized by EMC Isilon. I can connect to HDFS from the HDFS Dremio plugin, but it gives me an error when connecting to hive. Have any of you stumbled upon such a situation?
Thanks in advance :slight_smile:

@mdouda

There should be a copy icon right next to the Hive table name (on Dremio UI), click that. Then “New Query” and run the below query

select * from

Then click on jobs and click on the failed job for this query and on the right side scroll down to download the query profile. Send us the zip file

Thanks
@balaji.ramaswamy

Hi, Balaji.

Here’s the zip for a failed query.
I hope you can detect my problem better than I can.

Thank you for your attention :slight_smile:
5915140c-67be-48ab-a6a6-3fdfa2cb8011.zip (5.1 KB)

I see “connection refused” from “tqbigdprodldremiomr01” to the Namenode (isilonscprodhadoop.sari.local) on 8020. Now I also see you have 3 coordinators. Currently we do not recommend more than one active coordinator (not HA). Please use only one master-coordinator and validate again. Whichever host you are going to choose as your coordinator, make sure you can telnet to the NN host on 8020

“node”: [
{
“name”: “tqbigdprodldremioexor02”,
“role”: “executor”
},
{
“name”: “tqbigdprodldremioexor03”,
“role”: “executor”
},
{
“name”: “tqbigdprodldremioexor01”,
“role”: “executor”
},
{
“name”: “tqbigdprodldremiocoor02”,
“role”: “coordinator”
},
{
“name”: “tqbigdprodldremiomr01”,
“role”: “coordinator”
},
{
“name”: “tqbigdprodldremiocoor01”,
“role”: “coordinator”

Hello again, Balaji, and thank you for your quick response.
I tried to run the query again after removing the additional coordinators with no success. The error persists.
I checked with telnet and the connection is successful. Actually, I have HDFS connected with this same address and it does not complain, but Hive seems to not like it.
I followed the steps for configuring Hive according to Dremio documentation, modifying the needed files from Ambari.
Do you believe this could be fixed by upgrading Dremio to enterprise? It’s a community edition we have installed and we have the idea that maybe this may help.

Can you please share your hdfs-site.xml?

Gladly.

hdfs-site.zip.zip (1.3 KB)
Sorry for the format. Would not let me upload .xml files.