Hey Jeff, The Cloudera Training Virtual Machine<http://www.cloudera.com/hadoop-training-virtual-machine> contains two conflicting versions of libthrift.jar and libfb303.jar. The ones in the /usr/lib/hadoop/lib is an older version than /usr/lib/hive/lib. This causes the Hive server connectivity issue (see detailed explanation in Hive-875). I recommend upgrade the jar files in the hadoop's lib directory to the same as the ones in the hive directory.
Thanks, Ning On Oct 14, 2009, at 3:57 PM, Ning Zhang wrote: BTW, the Hive server standalone mode unit test is failing in the latest revision. This will affect the ODBC connection as well. I'm working on a fix (HIVE-875). If you have a working revision, stick with it first before this JIRA is fixed. Thanks, Ning On Oct 14, 2009, at 3:10 PM, Jeff Hammerbacher wrote: Hey Ning, Thanks for the quick reply. I'll do some more investigation on my end to see which APIs the GUI is using and which are not supported. The GUI, btw, is http://www.sqldbx.com. Regards, Jeff On Wed, Oct 14, 2009 at 2:03 PM, Ning Zhang <[email protected]<mailto:[email protected]>> wrote: Hi Jeff, The unixODBC implementation supports ODBC 3.0 standard (see http://en.wikipedia.org/wiki/Open_Database_Connectivity#UnixODBC ), although the version number of unixODBC on which we had a patch is 2.x. However, patches in Hive-187 only implemented a subset of ODBC 3.0 interface (see the end of the page http://wiki.apache.org/hadoop/Hive/HiveODBC ). If the software requires some some API that are not supported currently, it won't work. Could you let us know which ODBC functions are required by the software? Also note that even though the all required functions in the API are supported, it does not mean the software can integrate with Hive seamlessly. For example, the SQLExecute() function is supported by Hive's ODBC driver, but it doesn't mean any SQL queries passed by SQLExecute() can be executed by Hive. We are working on extending HiveQL to support 3rd party applications access Hive through ODBC drivers. In this respect, it would be helpful to know which SQL features the software are using so that we can put them in the roadmap. Thanks, Ning On Oct 14, 2009, at 1:31 PM, Jeff Hammerbacher wrote: Hey, I started working with the ODBC patch today ( https://issues.apache.org/jira/browse/HIVE-187) and noticed that the ODBC driver off of which the work is based (unixODBC) is 2.x, not 3.0. One of the database GUIs we're working with here requires ODBC 3.0. What was the reason for using unixODBC, and how much work would moving to a new ODBC driver require? I'm pretty out of my depth in this domain, so any help is appreciated. Regards, Jeff
