HiveServer

Skip to end of metadata

Page restrictions apply

Added by Confluence Administrator, last edited by Carl Steinbach on Apr 17, 2012

(view change)

Go to start of metadata

WARNING!

HiveServer can not handle concurrent requests from more than one client. This is actually a limitation imposed by the Thrift interface that HiveServer exports, and cant be resolved by modifying the HiveServer code. Were currently working on a rewrite of HiveServer that addresses these problems. Progress on this issue is being tracked in HIVE-2935.

Hive server and clients communicates through Thrift and FB303 services. In some distributions, both the Hadoop and Hive distributions have different versions of libthrift.jar and libfb303.jar. If they are incompatible, it may cause Thrift connection error when running the unit test on standalone mode. The solution is to remove the Hadoops version of libthrift.jar and libfb303.jar.

Once Hive has been built using steps in Getting Started, the thrift server can be started by running the following:

$ build/dist/bin/hive --service hiveserver --help

usage HIVE_PORT=xxxx ./hive --service hiveserver

HIVE_PORT : Specify the server port

$ bin/hive --service hiveserver

After starting the server, to test if the server is working well, run the hiveserver and jdbc tests in standalone mode. The HIVE_PORT is assumed to be 10000 on localhost for this case.

$ ant test -Dtestcase=TestJdbcDriver -Dstandalone=true

$ ant test -Dtestcase=TestHiveServer -Dstandalone=true

The service supports clients in multiple languages. For more details see Hive Client

Labels parameters