livy can't list databases

[lk_hadoop]

hi,all:
   I'm trying livy0.4 with spark2.1
   curl -H "Content-type: application/json" -X POST 
http://kafka02:8998/sessions -d '{"kind": "spark"}' | python -m json.tool
   curl -H "Content-type: application/json" -X POST 
http://kafka02:8998/sessions/0/statements -d '{"code": "spark.sql(\"show 
databases\").show"}' | python -m json.tool
   curl  http://kafka02:8998/sessions/0/statements/0 | python -m json.tool
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   235  100   235    0     0  15959      0 --:--:-- --:--:-- --:--:-- 16785
{
    "code": "spark.sql(\"show databases\").show",
    "id": 0,
    "output": {
        "data": {
            "text/plain": "+------------+\n|databaseName|\n+------------+\n|    
 default|\n+------------+"
        },
        "execution_count": 0,
        "status": "ok"
    },
    "progress": 1.0,
    "state": "available"
}

It looks like can't read the metadata,I have config livy with SPARK_HOME,and 
run under yarn model,the hive-site.xml also cp to SPARK_HOME/conf/.
but't when I use spark-shell:


scala> spark.sql("show databases").show
+-------------+
| databaseName|
+-------------+
|      default|
| tpcds_carbon|
|tpcds_carbon2|
| tpcds_indexr|
|tpcds_parquet|
| tpcds_source|
+-------------+



2017-11-22


lk_hadoop