You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@batchee.apache.org by "Hudson (JIRA)" <ji...@apache.org> on 2016/05/15 20:24:12 UTC
[jira] [Commented] (BATCHEE-88) rollback fails if the connection
already got closed by the container
[ https://issues.apache.org/jira/browse/BATCHEE-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15283948#comment-15283948 ]
Hudson commented on BATCHEE-88:
-------------------------------
SUCCESS: Integrated in batchee #610 (See [https://builds.apache.org/job/batchee/610/])
BATCHEE-88 unit test which shows the broken TX error handling (struberg: rev 4313c77012a8ed43de3b0b947f9857711b11f65a)
* jbatch/src/main/java/org/apache/batchee/spi/TransactionManagerAdapter.java
* integration-tests/transaction/src/test/java/org/apache/batchee/its/transaction/TxErrorReader.java
* pom.xml
* integration-tests/transaction/src/test/java/org/apache/batchee/its/transaction/TxErrorTest.java
* integration-tests/transaction/pom.xml
* integration-tests/transaction/src/test/java/org/apache/batchee/its/transaction/TxErrorWriter1.java
* integration-tests/transaction/src/test/java/org/apache/batchee/its/transaction/MockTransactionManagerService.java
* integration-tests/transaction/src/test/resources/batchee.properties
* integration-tests/pom.xml
* integration-tests/transaction/src/test/resources/META-INF/batch-jobs/txtest1.xml
BATCHEE-88 improve handling in case of error during step commit (struberg: rev 7a74df4bc9295377ab35ea53eacb8204698cd425)
* integration-tests/transaction/src/test/resources/META-INF/batch-jobs/txtest1.xml
* jbatch/src/main/java/org/apache/batchee/container/impl/controller/chunk/CheckpointManager.java
* integration-tests/transaction/src/test/java/org/apache/batchee/its/transaction/TxErrorTest.java
* jbatch/src/main/java/org/apache/batchee/container/impl/controller/chunk/ChunkStepController.java
* jbatch/src/main/java/org/apache/batchee/container/impl/controller/BaseStepController.java
> rollback fails if the connection already got closed by the container
> --------------------------------------------------------------------
>
> Key: BATCHEE-88
> URL: https://issues.apache.org/jira/browse/BATCHEE-88
> Project: BatchEE
> Issue Type: Bug
> Components: jbatch-core
> Affects Versions: 0.3-incubating
> Reporter: Mark Struberg
> Assignee: Mark Struberg
> Fix For: 0.4-incubating
>
> Attachments: BATCHEE-88.patch
>
>
> We get the following Exception in some cases. Happens if our Oracle closes the connection alltogether after a constraint problem in the db.
> {code}
> 2016-01-15 09:16:29,805 [@Asynchronous myapp-batches - 1] noSessionActive noUserId ERROR controller.BaseStepController No transaction associated with current thread
> java.lang.IllegalStateException: No transaction associated with current thread
> at org.apache.geronimo.transaction.manager.TransactionManagerImpl.setRollbackOnly(TransactionManagerImpl.java:241)
> at org.apache.openejb.core.CoreUserTransaction.setRollbackOnly(CoreUserTransaction.java:92)
> at org.apache.batchee.tools.services.thread.UserTransactionTransactionService$UserTransactionTxAdapter.setRollbackOnly(UserTransactionTransactionService.java:102)
> at org.apache.batchee.container.impl.controller.chunk.ChunkStepController.rollback(ChunkStepController.java:650)
> at org.apache.batchee.container.impl.controller.chunk.ChunkStepController.invokeChunk(ChunkStepController.java:632)
> at org.apache.batchee.container.impl.controller.chunk.ChunkStepController.invokeCoreStep(ChunkStepController.java:666)
> at org.apache.batchee.container.impl.controller.BaseStepController.execute(BaseStepController.java:152)
> at org.apache.batchee.container.impl.controller.ExecutionTransitioner.doExecutionLoop(ExecutionTransitioner.java:106)
> at org.apache.batchee.container.impl.controller.JobThreadRootController.originateExecutionOnThread(JobThreadRootController.java:110)
> at org.apache.batchee.container.util.BatchWorkUnit.run(BatchWorkUnit.java:62)
> {code}
> Any rollback exception should get catched away and the execution should continue normally imo.
> Any thoughts?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)