Hi,

I set up a standalone spark node, with 1 master, 1 worker and 1 instance of
spark jdbc server.
With that i was able to create 3 external hive tables (connected to
elasticsearch).

Using SQL workbench <http://www.sql-workbench.net/>, with connection
string jdbc:hive2://host:10000
and driver class org.apache.hive.jdbc.HiveDriver, i am able to succesfully
run various queries.

Next, i tried to connect apache drill 1.8 to the spark jdbc server.
I managed to succesfully add a jdbc storage plugin, configured as follows:

{
  "type": "jdbc",
  "driver": "org.apache.hive.jdbc.HiveDriver",
  "url": "jdbc:hive2://host:10000",
  "username": null,
  "password": null,
  "enabled": true
}

Ofcourse, all required jars were put in the \jars\3rdparty folder, and
indeed the storage plugin was added succesfully via with drill web console.

Running "show schemas" returns the new added storage plugin and workspace.
However, "show tables" return no tables.
Executing any query that worked succesfully with the SQL workbench results
in the following error:
org.apache.drill.common.exceptions.UserRemoteException: VALIDATION ERROR:
>From line 1, column 15 to line 1, column 23: Table 'sparkhive.customer' not
found SQL Query null

As a last resort, i attempted to to create a second storage plugin, a hive
plugin, in the hopes that this will help drill to get schema information of
tables:

{
  "type": "hive",
  "enabled": true,
  "configProps": {
  "hive.metastore.uris":
"file:///D:/spark-2.0.2-bin-hadoop2.7/bin/metastore_db",
  "javax.jdo.option.ConnectionDriverName":
"org.apache.hive.jdbc.HiveDriver",
  "javax.jdo.option.ConnectionURL": "jdbc:hive2://host:10000",
  "hive.metastore.warehouse.dir":
"file:///D:/spark-2.0.2-bin-hadoop2.7/bin/spark-warehouse",
  "fs.default.name": "file:///",
  "hive.metastore.sasl.enabled": "false"
  }
}

but this didnt work.
The plugin cannot be saved succesfully.

What could the be the problem of this setup?

Thanks,
Roman Margolis

Reply via email to