Douglas created HIVE-5296: ----------------------------- Summary: Memory leak: OOM Error after multiple open/closed JDBC connections. Key: HIVE-5296 URL: https://issues.apache.org/jira/browse/HIVE-5296 Project: Hive Issue Type: Bug Components: HiveServer2 Affects Versions: 0.12.0 Environment: Hive 0.12.0, Hadoop 1.1.2, Debian. Reporter: Douglas Fix For: 0.12.0
This error seems to relate to https://issues.apache.org/jira/browse/HIVE-3481 However, on inspection of the related patch and my built version of Hive (patch carried forward to 0.12.0), I am still seeing the described behaviour. Multiple connections to Hiveserver2, all of which are closed and disposed of properly show the Java heap size to grow extremely quickly. This issue can be recreated using the following code {code} import java.sql.DriverManager; import java.sql.Connection; import java.sql.ResultSet; import java.sql.SQLException; import java.sql.Statement; import java.util.Properties; import org.apache.hive.service.cli.HiveSQLException; import org.apache.log4j.Logger; /* * Class which encapsulates the lifecycle of a query or statement. * Provides functionality which allows you to create a connection */ public class HiveClient { Connection con; Logger logger; private static String driverName = "org.apache.hive.jdbc.HiveDriver"; private String db; public HiveClient(String db) { logger = Logger.getLogger(HiveClient.class); this.db=db; try{ Class.forName(driverName); }catch(ClassNotFoundException e){ logger.info("Can't find Hive driver"); } String hiveHost = GlimmerServer.config.getString("hive/host"); String hivePort = GlimmerServer.config.getString("hive/port"); String connectionString = "jdbc:hive2://"+hiveHost+":"+hivePort +"/default"; logger.info(String.format("Attempting to connect to %s",connectionString)); try{ con = DriverManager.getConnection(connectionString,"",""); }catch(Exception e){ logger.error("Problem instantiating the connection"+e.getMessage()); } } public int update(String query) { Integer res = 0; Statement stmt = null; try{ stmt = con.createStatement(); String switchdb = "USE "+db; logger.info(switchdb); stmt.executeUpdate(switchdb); logger.info(query); res = stmt.executeUpdate(query); logger.info("Query passed to server"); stmt.close(); }catch(HiveSQLException e){ logger.info(String.format("HiveSQLException thrown, this can be valid, " + "but check the error: %s from the query %s",query,e.toString())); }catch(SQLException e){ logger.error(String.format("Unable to execute query SQLException %s. Error: %s",query,e)); }catch(Exception e){ logger.error(String.format("Unable to execute query %s. Error: %s",query,e)); } if(stmt!=null) try{ stmt.close(); }catch(SQLException e){ logger.error("Cannot close the statment, potentially memory leak "+e); } return res; } public void close() { if(con!=null){ try { con.close(); } catch (SQLException e) { logger.info("Problem closing connection "+e); } } } } {code} And by creating and closing many HiveClient objects. The heap space used by the hiveserver2 runjar process is seen to increase extremely quickly, without such space being released. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira