livy-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From vanzin <...@git.apache.org>
Subject [GitHub] incubator-livy issue #118: [LIVY-503] Separate thrift server session code in...
Date Tue, 27 Nov 2018 18:46:38 GMT
Github user vanzin commented on the issue:

    https://github.com/apache/incubator-livy/pull/118
  
    BTW is it just me or there are actually no integration tests for the thrift server? There
are unit tests that run the thrift server, but that's different. If that's the case, it's
another thing that should be added before this is released...
    
    I fixed a Kryo issue. I can now execute things. "select"-style queries work, other queries
(like "create table") also work but there's an error while rendering the result in beeline.
I compared the code to the previous in that area and I don't really see a difference (the
code that translates Spark's json schema to a Hive schema is the same). The statement itself
works, though, you just see an ugly exception in beeline.
    
    e.g.
    
    ```
    : jdbc:hive2://vanzin-c5-1.gce.cloudera.com:> drop table test;
    java.lang.NullPointerException
            at org.apache.hive.service.cli.ColumnBasedSet.<init>(ColumnBasedSet.java:50)
            at org.apache.hive.service.cli.RowSetFactory.create(RowSetFactory.java:37)
    ```
    
    I'll take another look though.


---

Mime
View raw message