[HIVE-20927] The beeline has a dead cycle, resulting in high CPU.

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.3.0
• Component/s: Clients
• Assignee: Unassigned
• Reporter: yinghua_zh
• Created: 16/Nov/18 05:21
• Updated: 16/Nov/18 05:47

Description

When beeline executes the command and places it for a period of time, it is found that the CPU of the beeline process is very high. It is the code that has a dead loop.The stack is as follows:

"Thread-4" #34 daemon prio=5 os_prio=0 tid=0x00007fbd6636d800 nid=0x162be sleeping[0x00007fbd2e042000]
java.lang.Thread.State: RUNNABLE
at java.lang.Throwable.fillInStackTrace(Native Method)
at java.lang.Throwable.fillInStackTrace(Throwable.java:783)

• locked <0x000000076edb39d8> (a java.sql.SQLException)
  at java.lang.Throwable.<init>(Throwable.java:265)
  at java.lang.Exception.<init>(Exception.java:66)
  at java.sql.SQLException.<init>(SQLException.java:119)
  at org.apache.hive.jdbc.HiveStatement.checkConnection(HiveStatement.java:377)
  at org.apache.hive.jdbc.HiveStatement.getQueryLog(HiveStatement.java:865)
  at org.apache.hive.jdbc.HiveStatement.getQueryLog(HiveStatement.java:849)
  at org.apache.hive.beeline.Commands$LogRunnable.updateQueryLog(Commands.java:1213)
  at org.apache.hive.beeline.Commands$LogRunnable.run(Commands.java:1221)
  at java.lang.Thread.run(Thread.java:745)

 

The dead cycle code is as is as follows:

@Override public void run() {
while (hiveStatement.hasMoreLogs()) {
try

{ updateQueryLog(); Thread.sleep(queryProgressInterval); }

catch (SQLException e)

{ commands.error(new SQLWarning(e)); }

catch (InterruptedException e)

{ commands.debug("Getting log thread is interrupted, since query is done!"); commands.showRemainingLogsIfAny(hiveStatement); }

}
}
}

   

 

when the mehtod (checkConnection("getQueryLog");)occurs exception,the condition "hiveStatement.hasMoreLogs()" is still true,exist the dead lop,but the thread sleep code is not be run,so the CPU is high!