openjpa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Albert Lee (JIRA)" <j...@apache.org>
Subject [jira] Commented: (OPENJPA-1753) TestMixedLockManagerLockPermutation: Timing issue determines the pass/no-pass of the test case
Date Thu, 12 Aug 2010 20:40:15 GMT

    [ https://issues.apache.org/jira/browse/OPENJPA-1753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12897939#action_12897939
] 

Albert Lee commented on OPENJPA-1753:
-------------------------------------

Need to set postgresql.conf property max_connection >= 150, otherwise will get exception
like:

TestMixedLockManagerLockPermutation.testLockWriteRead  View test details   (org.apache.openjpa.persistence.lockmgr)
    <openjpa-2.1.0-SNAPSHOT-rexported fatal store error> org.apache.openjpa.util.StoreException:
Cannot create PoolableConnectionFactory (FATAL: sorry, too many clients already)
    at org.apache.openjpa.jdbc.sql.DBDictionaryFactory.newDBDictionary(DBDictionaryFactory.java:102)
    at org.apache.openjpa.jdbc.conf.JDBCConfigurationImpl.getDBDictionaryInstance(JDBCConfigurationImpl.java:599)


> TestMixedLockManagerLockPermutation: Timing issue determines the pass/no-pass of the
test case
> ----------------------------------------------------------------------------------------------
>
>                 Key: OPENJPA-1753
>                 URL: https://issues.apache.org/jira/browse/OPENJPA-1753
>             Project: OpenJPA
>          Issue Type: Bug
>    Affects Versions: 2.1.0
>            Reporter: Fay Wang
>            Priority: Minor
>             Fix For: 2.1.0
>
>         Attachments: trace_fail.txt, trace_success.txt
>
>
> This test case spawns three threads and the pass/no pass of the test cases depends on
the timing. 
> 	One example is TestMixedLockManagerFindPermutation.testFindReadPessimisticRead: 
>         commonFindTest(
>             "testFind(Read,Commit/PessimisticRead,Rollback)",
>             LockModeType.READ, Act.CommitTx, 1, null,
>             LockModeType.PESSIMISTIC_READ, Act.RollbackTx, 1, null);
> 	If the find by the second child [Thread-8] is executed before the jdbc commit of the
update statement by the first child [Thread-7], the test will fail. Attached are the trace
file for the successful run and failing run. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message