Hi,

I have ipc.client.connection.maxidletime set to default (10 sec). The hive-server2 threads (and tcp connections to zookeeper) stay active forever. I guess these issues might be related but they are not the same.

Regards,
Dima Fadeyev

El 05/05/14 03:47, Shengjun Xin escribió:
Is it same with https://issues.apache.org/jira/browse/HIVE-6866 ?


On Tue, Apr 29, 2014 at 8:26 PM, Dima Fadeyev <dfade...@pragsis.com <mailto:dfade...@pragsis.com>> wrote:

    Hi, Chinna. Thanks for your reply.

    Yes, modifying code solves the problem. This is what my code looks
    like (a piece of it):

        Connection con =
    DriverManager.getConnection("jdbc:hive2://localhost:10000/default", "hive",
    "hive");
        Statement stmt = con.createStatement();
        String tableName = "testHiveDriverTable";
        stmt.execute("drop table if exists " + tableName);
        //stmt.close();

    When I uncomment the last line, the number of hive-server2 threads
    doesn't keep incrementing to infinity. However I'm investigating
    the issue where the code is not really my code. Is there a way to
    correct this behavior from within hive-server2 without changing
    the client's code?

    El 29/04/14 14:05, Chinna Rao Lalam escribió:
    Hi,

     In your code if more connections and statements are created?. If
    so, closed those connections?
     After use close unused connections and statements.


    Hope It Helps,
    Chinna


    On Tue, Apr 29, 2014 at 3:47 PM, Dima Fadeyev
    <dfade...@pragsis.com <mailto:dfade...@pragsis.com>> wrote:

        Hello everyone,

        When I run a jdbc example from
        
https://cwiki.apache.org/confluence/display/Hive/HiveServer2+Clients#HiveServer2Clients-JDBCClientSampleCode
        against my hive server, the number of hive-server2 threads
        increments. If I execute it long enough I either start seeing
        exceptions

        Exception in thread "main" java.sql.SQLException:
        org.apache.thrift.TApplicationException: Internal error
        processing ExecuteStatement
            at
        org.apache.hive.jdbc.HiveStatement.execute(HiveStatement.java:203)
            at HiveJdbcClient.main(HiveJdbcClient.java:24)
        Caused by: org.apache.thrift.TApplicationException: Internal
        error processing ExecuteStatement
            at
        
org.apache.thrift.TApplicationException.read(TApplicationException.java:108)
            at
        org.apache.thrift.TServiceClient.receiveBase(TServiceClient.java:71)
            at
        
org.apache.hive.service.cli.thrift.TCLIService$Client.recv_ExecuteStatement(TCLIService.java:213)
            at
        
org.apache.hive.service.cli.thrift.TCLIService$Client.ExecuteStatement(TCLIService.java:200)
            at
        org.apache.hive.jdbc.HiveStatement.execute(HiveStatement.java:197)
            ... 1 more

        or I bump into a limit of connections with Zookeeper (each
        hive-server2 thread maintains a connection with Zookeeper. I
        have hive.support.concurrency enabled).

        In either case I can't connect to hive server after that.

        I've tried this on hive 0.10 (cdh 4.4) and hive 0.12 (cdh 5.0
        and hdp 2.0.6) with same results.

        Please, could anyone help me resolve this.
        Thanks in advance.





-- Hope It Helps,
    Chinna




--
Regards
Shengjun

Reply via email to