Mailing-List: contact issues-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Thu, 30 Jul 2015 13:07:04 +0000 (UTC)
From: "shiqian.huang (JIRA)" <jira@apache.org>
To: issues@hive.apache.org
Message-ID: <JIRA.12850227.1438259974000.330104.1438261624524@Atlassian.JIRA>
In-Reply-To: <JIRA.12850227.1438259974000@Atlassian.JIRA>
References: <JIRA.12850227.1438259974000@Atlassian.JIRA>
 <JIRA.12850227.1438259974079@arcas>
Subject: [jira] [Updated] (HIVE-11411) Transaction lock time out when can't
 tracking job progress
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/HIVE-11411?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

shiqian.huang updated HIVE-11411:
---------------------------------
    Description: 
in method progress(ExecDriverTaskHandle th) of class HadoopJobExecHelper, hive lock heartbeating progress works with job tracking progress. 
such like 
      heartbeater.heartbeat();

      if (initializing && rj.getJobState() == JobStatus.PREP) {
        // No reason to poll untill the job is initialized
        continue;
      } else {
        // By now the job is initialized so no reason to do
        // rj.getJobState() again and we do not want to do an extra RPC call
        initializing = false;
      }
When job tracking progress got any exception in  rj.getJobState() == JobStatus.PREP, will bring  NoSuchLockException(hive client & exception message:No record of lock could be found, may have timed out) to big query job finally. 

  was:
in method progress(ExecDriverTaskHandle th) of class HadoopJobExecHelper, hive lock heartbeating progress works with job tracking progress. 
such like 
      heartbeater.heartbeat();

      if (initializing && rj.getJobState() == JobStatus.PREP) {
        // No reason to poll untill the job is initialized
        continue;
      } else {
        // By now the job is initialized so no reason to do
        // rj.getJobState() again and we do not want to do an extra RPC call
        initializing = false;
      }
When job tracking progress got any exception in  rj.getJobState() == JobStatus.PREP, will bring lock time out exception to big query job finally.  


> Transaction lock time out when can't tracking job progress
> ----------------------------------------------------------
>
>                 Key: HIVE-11411
>                 URL: https://issues.apache.org/jira/browse/HIVE-11411
>             Project: Hive
>          Issue Type: Wish
>    Affects Versions: 1.2.0
>            Reporter: shiqian.huang
>            Priority: Minor
>
> in method progress(ExecDriverTaskHandle th) of class HadoopJobExecHelper, hive lock heartbeating progress works with job tracking progress. 
> such like 
>       heartbeater.heartbeat();
>       if (initializing && rj.getJobState() == JobStatus.PREP) {
>         // No reason to poll untill the job is initialized
>         continue;
>       } else {
>         // By now the job is initialized so no reason to do
>         // rj.getJobState() again and we do not want to do an extra RPC call
>         initializing = false;
>       }
> When job tracking progress got any exception in  rj.getJobState() == JobStatus.PREP, will bring  NoSuchLockException(hive client & exception message:No record of lock could be found, may have timed out) to big query job finally. 


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)