[ 
https://issues.apache.org/jira/browse/TRAFODION-2724?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16139550#comment-16139550
 ] 

ASF GitHub Bot commented on TRAFODION-2724:
-------------------------------------------

Github user sandhyasun commented on a diff in the pull request:

    
https://github.com/apache/incubator-trafodion/pull/1216#discussion_r134926173
  
    --- Diff: core/sql/exp/ExpLOBaccess.cpp ---
    @@ -1970,7 +1970,14 @@ Ex_Lob_Error ExLob::readCursorData(char *tgt, Int64 
tgtSize, cursor_t &cursor, I
                   continue;
              }
           }
    -      
    +      if (operLen == tgtSize)
    +        {
    +          //The cursor still has data available but wait until redrive 
since
    +          //we have fetched all the data for this extract operation
    --- End diff --
    
    The cursor on the full lob data indicates that there is data left. But the 
extract operation can retrieve justt a portion of the total. The state is 
maintained in the cursor and will get updated with each "execute" of the 
extract statement issued by caller of the extract statement.  Agree this 
comment is confusing. I can just say "return the  data to caller". 


> JDBC LOB tests show symptoms of a leaked statement handle
> ---------------------------------------------------------
>
>                 Key: TRAFODION-2724
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-2724
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: sql-exe
>    Affects Versions: 2.2-incubating
>            Reporter: Sandhya Sundaresan
>            Assignee: Sandhya Sundaresan
>             Fix For: 2.2-incubating
>
>
> The number of opens for any mxosrvr is limited to 13 due to a recent change. 
> This exposed an issue where we were not closing some internal statements 
> during an extract statement . JDBC driver uses the SQL extract statment to 
> implement the stream API. This exposed the issue.
> The test failed with this error:
> > 2017-08-16 01:12:27 Running testBatchInsertFKNotExist......
> > 2017-08-16 01:12:56 Batch Update Failed, See next exception for 
> > details
> > 2017-08-16 01:12:56 *** ERROR[8103] The operation is prevented by 
> > referential integrity constraint
> > TRAFODION.T4QA.BATCH_TEST_TABLE_FK_911645628_4551 on table 
> > TRAFODION.T4QA.BATCH_TEST_TABLE_FK.
> > 2017-08-16 01:13:11 *** ERROR[2026] Server Process $ZSM0 has reached 
> > allowed depth for nowait operation from the process 0,15846.
> > [2017-08-16 01:13:11]
> > 
> Debugging and dumping hte open and fetch statements in the context at that 
> point showed this pattern :
> There were around 13 statements in OPEN or FETCH statement with the 
> > source string "select dataOffset, chunkLen, stringParam from 
> > table(ghost table
> > "TRAFODION"."T4QA"."LOBDescChunks__08574991311373262837_0002")
> > where descPartnKey = and descSyskey = "
> > 
> > These internal statements should be closed for the jdbc_test_cdh to 
> > pass or comment out the tests that triggers these internal statements 
> > for time being.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to