Radoop Connection Failed Several Times

data123

My Radoop test repeatedly fails. Any clues?

 

[Feb 4, 2018 3:40:54 PM] SEVERE: java.sql.SQLException: Could not open client transport with JDBC Uri: jdbc:hive2://x.x.x.x:10000/default;radoop.proxy.source=123%25%25%25local%25%25%25123;radoop.proxy.name=Hadoop_Proxy: java.net.ConnectException: Connection timed out: connect
[Feb 4, 2018 3:40:54 PM] SEVERE: Connection Error for Driver "HiveDriverShim (radoop_hive_0.13.0): org.apache.hive.jdbc.HiveDriver@538c823c" with URL "radoop_hive_0.13.0jdbc:hive2://x.x.x.x:10000/default;radoop.proxy.source=123%25%25%25local%25%25%25123;radoop.proxy.name=Hadoop_Proxy", username("xxxx"), password("***************") Message: Could not open client transport with JDBC Uri: jdbc:hive2://x.x.x.x:10000/default;radoop.proxy.source=123%25%25%25local%25%25%25123;radoop.proxy.name=Hadoop_Proxy: java.net.ConnectException: Connection timed out: connect
[Feb 4, 2018 3:40:54 PM] SEVERE: Test failed: Hive connection
[Feb 4, 2018 3:40:54 PM] SEVERE: Connection test for 'myHive2' failed.

phellinger

Hi,

 

is this a connection that used to work and it does not now? Or is this the first try?

 

A Radoop Proxy is used: you can find the log of this service at the radoop-proxy/logs/ subfolder of the Server installation directory. Does it show the connection request, or the test does not even get there? Depending on this, you should either check the connection to the Radoop Proxy from your client (Studo), or check the connection from the Proxy to the HiveServer2 instance (if the Proxy does contain log entries). Using tools like ping, telnet, or nc can check if the address or the specific port is accessible. 

 

If needed, timeout values can be increased (usually 30 seconds should be enough, though): https://docs.rapidminer.com/latest/radoop/troubleshooting/connection-errors.html

 

Best,

Peter

data123

Hi Peter,

Yes. This is happening at first attempt.

My connection is:

RM Studio -> RM Server (Radoop Proxy) -> Hadoop Cluster

Ping test from RM server to hadoop name node = succesful

Ping test from RM studio to RM server = succesful

Increased timeout to 60 from 30 but still fails.

Am I suppose to see my radoop connection under 'add connection' in server?

 

phellinger

Hi,

 

You should only see the Radoop Proxy connection there only if Radoop and its license is installed on Server as well (see instructions at docs.rapidminer.com/radoop; you can check success in Administration->Operators&Extensions).

That is only necessary, if you want to run Radoop processes on Server as well. (Using the Proxy does not require that.)

 

The "java.net.ConnectException: Connection timed out: connect" error could mainly be caused by either a firewall blocking somethin, or the service is not running (on that port), or some routing problem.

The HiveServer2 port can be checked in Ambari or Cloudera Manager or netstat on the cluster node depending on what distribution this is. Depending on the Transport mode (hive.server2.transport.mode), the default is 10000 (binary) or 10001 (http). But, actually, if it is a closed port, "Connection refused" is the expected error. Timeout indicates no response. The HiveServer2 health can be checkd. And its log can be checked for this connection attempt (Radoop runs a "SHOW TABLES" command as a test).

 

Best,

Peter