A jberet user reported JBERET-29<https://issues.jboss.org/browse/JBERET-29>(Foreign key constraint step_execution_jobexecutionid_fkey fails when using Postgresql on WildFly, and we are trying to fix it by jberet 1.0.0.Final. The problem happens when user app starts a job within a transaction (e.g., CMT EJB), jberet inserts JobExecution into database (thread 1 & transaction 1), and then spawn a jberet-batch thread to run the job (thead 2 & transaction 2). Sometimes T2 tries to access db before T1 is committed, hence the error reported by the user. What's the common approach for solving this kind of problem? I suppose other WildFly components may also have this issue and probably already solved. Using transaction synchronization is a cleaner solution than polling db, but I'm not sure about its full implication. Ideally, I don't want to use system level JTA API like TransactionManager or Synchronization in jberet proper, but probably we can implement it in WildFly jberet integration. Thanks, Cheng _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed.

What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not.

Stuart On Tue, Feb 4, 2014 at 4:49 PM, Cheng Fang <cfang(a)redhat.com <mailto:cfang@redhat.com>> wrote: A jberet user reported JBERET-29 <https://issues.jboss.org/browse/JBERET-29> (Foreign key constraint step_execution_jobexecutionid_fkey fails when using Postgresql on WildFly, and we are trying to fix it by jberet 1.0.0.Final. The problem happens when user app starts a job within a transaction (e.g., CMT EJB), jberet inserts JobExecution into database (thread 1 & transaction 1), and then spawn a jberet-batch thread to run the job (thead 2 & transaction 2). Sometimes T2 tries to access db before T1 is committed, hence the error reported by the user. What's the common approach for solving this kind of problem? I suppose other WildFly components may also have this issue and probably already solved. Using transaction synchronization is a cleaner solution than polling db, but I'm not sure about its full implication. Ideally, I don't want to use system level JTA API like TransactionManager or Synchronization in jberet proper, but probably we can implement it in WildFly jberet integration. Thanks, Cheng _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org <mailto:wildfly-dev@lists.jboss.org> https://lists.jboss.org/mailman/listinfo/wildfly-dev

On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: > On 2/4/14, 9:57 AM, Stuart Douglas wrote: >> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >> > yes, we could implement it in wildfly-batch integration module. >> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. > The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: > > Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: > > Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn’t describe it, and the RI doesn’t do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door.

-- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

On 02/04/2014 02:42 PM, James R. Perkins wrote: > > On 02/04/2014 08:16 AM, Jason Greene wrote: >> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >> >>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>> I would use a transaction synchronization, so you don't spawn the >>>> other thread until the transaction is successfully committed. >>>> >>> yes, we could implement it in wildfly-batch integration module. >>>> What does the spec say about transactions? If a job is create in a >>>> thread that is part of a transaction and the transaction is rolled >>>> back should the job actually go ahead? Common sense would suggest >>>> not. >>> The transaction treatment in the batch spec is mostly around item >>> processing, not much on how it interacts with the transaction in >>> the running environment. The only place that it touches on Java EE >>> environment is section 9.7 Transactionality: >>> >>> Chunk type check points are transactional. The batch runtime uses >>> global transaction mode on the Java EE platform and local >>> transaction mode on the Java SE platform. Global transaction >>> timeout is configurable at step-level with a step-level property: >>> >>> Yes, I agree if the batch client side transaction is rolled back, >>> the job execution should not proceed. With the current jberet >>> impl, the job execution in this case will fail since the job >>> repository is not in good state, like in the above bug. If we have >>> transaction syncrhonization in place, then the job will not start >>> running till transaction 1 is committed. >> There is a consistency problem here though. If you expect the client >> side to rollback on transaction failure, then the in-memory job >> store should as well. IMO before committing to such a big feature, I >> would recommend looking at what the RI does here. If the spec >> doesn’t describe it, and the RI doesn’t do it, then we should avoid >> investing time on it at least right now where we really need to get >> WF8 out the door. > I don't see in the spec where it requires any kind of transaction around > a job repository. In fact the spec states "Note the implementation of > the job repository is outside the scope of this specification.". > > The RI does have a JDBC repository, but it doesn't insert anything into > the tables in a transaction. > > If we're only seeing this in PostgreSQL and a workaround with putting > JobOperator.start() outside a transaction works, I would suggest that's > okay for now. I do agree it needs to be fixed, but we might want to look > at how we're handling transaction in JBeret as a whole. The RI, not that > I want to model anything after it, uses it's own > TransactionManagerAdapter. It might make sense for JBeret to use a > TransactionManager rather than a UserTransaction. Or put the ownness on > the SPI implementation of the BatchEnvironment to handle the > transactions. Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction?

>> >> -- >> Jason T. Greene >> WildFly Lead / JBoss EAP Platform Architect >> JBoss, a division of Red Hat >> >> >> _______________________________________________ >> wildfly-dev mailing list >> wildfly-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/wildfly-dev >

On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: > > On 02/04/2014 12:40 PM, Scott Marlow wrote: >> On 02/04/2014 02:42 PM, James R. Perkins wrote: >>> >>> On 02/04/2014 08:16 AM, Jason Greene wrote: >>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >>>> >>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>>>> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >>>>>> >>>>> yes, we could implement it in wildfly-batch integration module. >>>>>> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. >>>>> The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: >>>>> >>>>> Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: >>>>> >>>>> Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. >>>> There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn’t describe it, and the RI doesn’t do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door. >>> I don't see in the spec where it requires any kind of transaction around >>> a job repository. In fact the spec states "Note the implementation of >>> the job repository is outside the scope of this specification.". >>> >>> The RI does have a JDBC repository, but it doesn't insert anything into >>> the tables in a transaction. >>> >>> If we're only seeing this in PostgreSQL and a workaround with putting >>> JobOperator.start() outside a transaction works, I would suggest that's >>> okay for now. I do agree it needs to be fixed, but we might want to look >>> at how we're handling transaction in JBeret as a whole. The RI, not that >>> I want to model anything after it, uses it's own >>> TransactionManagerAdapter. It might make sense for JBeret to use a >>> TransactionManager rather than a UserTransaction. Or put the ownness on >>> the SPI implementation of the BatchEnvironment to handle the transactions. >> >> Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction? > No I'm just saying they need to invoke the JobOperator.start() outside a transaction. At least from my understand on the JIRA that seems to workaround the issue. I will admit to not fully looking into this in detail though ;) That would be silly :)

On Feb 4, 2014, at 3:13 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: > > On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: > >> >> On 02/04/2014 12:40 PM, Scott Marlow wrote: >>> On 02/04/2014 02:42 PM, James R. Perkins wrote: >>>> >>>> On 02/04/2014 08:16 AM, Jason Greene wrote: >>>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >>>>> >>>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>>>>> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >>>>>>> >>>>>> yes, we could implement it in wildfly-batch integration module. >>>>>>> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. >>>>>> The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: >>>>>> >>>>>> Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: >>>>>> >>>>>> Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. >>>>> There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn’t describe it, and the RI doesn’t do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door. >>>> I don't see in the spec where it requires any kind of transaction around >>>> a job repository. In fact the spec states "Note the implementation of >>>> the job repository is outside the scope of this specification.". >>>> >>>> The RI does have a JDBC repository, but it doesn't insert anything into >>>> the tables in a transaction. >>>> >>>> If we're only seeing this in PostgreSQL and a workaround with putting >>>> JobOperator.start() outside a transaction works, I would suggest that's >>>> okay for now. I do agree it needs to be fixed, but we might want to look >>>> at how we're handling transaction in JBeret as a whole. The RI, not that >>>> I want to model anything after it, uses it's own >>>> TransactionManagerAdapter. It might make sense for JBeret to use a >>>> TransactionManager rather than a UserTransaction. Or put the ownness on >>>> the SPI implementation of the BatchEnvironment to handle the transactions. >>> >>> Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction? >> No I'm just saying they need to invoke the JobOperator.start() outside a transaction. At least from my understand on the JIRA that seems to workaround the issue. I will admit to not fully looking into this in detail though ;) > > That would be silly :) Requiring a NOT_SUPPORTED method that is. It’s pretty easy for JBeret to isolate the transaction if it wanted to tx = TransactionManager.suspend() TransactionManager.begin() // write the record TransactionManager.commit() TransactionManager.resume(tx);

On Feb 4, 2014, at 3:51 PM, Radoslaw Rodak <rodakr(a)gmx.ch&gt; wrote: > Hi > > > Am 04.02.2014 um 22:16 schrieb Jason Greene <jason.greene(a)redhat.com&gt;: > >> >> On Feb 4, 2014, at 3:13 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: >> >>> >>> On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: >>> >>>> >>>> On 02/04/2014 12:40 PM, Scott Marlow wrote: >>>>> On 02/04/2014 02:42 PM, James R. Perkins wrote: >>>>>> >>>>>> On 02/04/2014 08:16 AM, Jason Greene wrote: >>>>>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >>>>>>> >>>>>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>>>>>>> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >>>>>>>>> >>>>>>>> yes, we could implement it in wildfly-batch integration module. >>>>>>>>> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. >>>>>>>> The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: >>>>>>>> >>>>>>>> Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: >>>>>>>> >>>>>>>> Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. >>>>>>> There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn't describe it, and the RI doesn't do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door. >>>>>> I don't see in the spec where it requires any kind of transaction around >>>>>> a job repository. In fact the spec states "Note the implementation of >>>>>> the job repository is outside the scope of this specification.". >>>>>> >>>>>> The RI does have a JDBC repository, but it doesn't insert anything into >>>>>> the tables in a transaction. >>>>>> >>>>>> If we're only seeing this in PostgreSQL and a workaround with putting >>>>>> JobOperator.start() outside a transaction works, I would suggest that's >>>>>> okay for now. I do agree it needs to be fixed, but we might want to look >>>>>> at how we're handling transaction in JBeret as a whole. The RI, not that >>>>>> I want to model anything after it, uses it's own >>>>>> TransactionManagerAdapter. It might make sense for JBeret to use a >>>>>> TransactionManager rather than a UserTransaction. Or put the ownness on >>>>>> the SPI implementation of the BatchEnvironment to handle the transactions. >>>>> >>>>> Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction? >>>> No I'm just saying they need to invoke the JobOperator.start() outside a transaction. At least from my understand on the JIRA that seems to workaround the issue. I will admit to not fully looking into this in detail though ;) >>> >>> That would be silly :) >> >> Requiring a NOT_SUPPORTED method that is. It's pretty easy for JBeret to isolate the transaction if it wanted to >> >> tx = TransactionManager.suspend() >> TransactionManager.begin() >> // write the record >> TransactionManager.commit() >> TransactionManager.resume(tx); >> > > What will happened to suspended Transaction when you get Exception on TransactionManager.commit() ? You put resume in a finally block. Just like RequiresNew effectively does. -- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

You're suggesting something more like TransactionManager.getTransaction().registerSynchronization(synchronization); Rather than the; TransactionManager.suspend(); try { doStuff(); } finally { TransactionManager.resume(); } Am I understanding that correctly? On 02/04/2014 10:32 PM, Stuart Douglas wrote: This all sounds like a very similar problem to what we already do with EJB timers. Timers are transactional, if you create or cancel a timer it does not take effect until the transaction commits. The way this is accomplished is two fold: - The data store is transactional (or semi-transactional really in the case of the file data store, as we did not develop a fully transactional file system just for this) - Timers are not actually started or cancelled until the afterComplete() synchronization runs. I think it would make sense for JBeret to basically do the same. I think it would be very surprising to the user if jobs they started in transactions that abort just proceed as normal. Stuart On Tue, Feb 4, 2014 at 11:53 PM, Jason Greene <jason.greene(a)redhat.com&gt;wrote: > > On Feb 4, 2014, at 3:51 PM, Radoslaw Rodak <rodakr(a)gmx.ch&gt; wrote: > > > Hi > > > > > > Am 04.02.2014 um 22:16 schrieb Jason Greene <jason.greene(a)redhat.com&gt;: > > > >> > >> On Feb 4, 2014, at 3:13 PM, Jason Greene <jason.greene(a)redhat.com&gt; > wrote: > >> > >>> > >>> On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; > wrote: > >>> > >>>> > >>>> On 02/04/2014 12:40 PM, Scott Marlow wrote: > >>>>> On 02/04/2014 02:42 PM, James R. Perkins wrote: > >>>>>> > >>>>>> On 02/04/2014 08:16 AM, Jason Greene wrote: > >>>>>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: > >>>>>>> > >>>>>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: > >>>>>>>>> I would use a transaction synchronization, so you don't spawn > the other thread until the transaction is successfully committed. > >>>>>>>>> > >>>>>>>> yes, we could implement it in wildfly-batch integration module. > >>>>>>>>> What does the spec say about transactions? If a job is create > in a thread that is part of a transaction and the transaction is rolled > back should the job actually go ahead? Common sense would suggest not. > >>>>>>>> The transaction treatment in the batch spec is mostly around > item processing, not much on how it interacts with the transaction in the > running environment. The only place that it touches on Java EE environment > is section 9.7 Transactionality: > >>>>>>>> > >>>>>>>> Chunk type check points are transactional. The batch runtime > uses global transaction mode on the Java EE platform and local transaction > mode on the Java SE platform. Global transaction timeout is configurable at > step-level with a step-level property: > >>>>>>>> > >>>>>>>> Yes, I agree if the batch client side transaction is rolled > back, the job execution should not proceed. With the current jberet impl, > the job execution in this case will fail since the job repository is not in > good state, like in the above bug. If we have transaction syncrhonization > in place, then the job will not start running till transaction 1 is > committed. > >>>>>>> There is a consistency problem here though. If you expect the > client side to rollback on transaction failure, then the in-memory job > store should as well. IMO before committing to such a big feature, I would > recommend looking at what the RI does here. If the spec doesn't describe > it, and the RI doesn't do it, then we should avoid investing time on it at > least right now where we really need to get WF8 out the door. > >>>>>> I don't see in the spec where it requires any kind of transaction > around > >>>>>> a job repository. In fact the spec states "Note the implementation > of > >>>>>> the job repository is outside the scope of this specification.". > >>>>>> > >>>>>> The RI does have a JDBC repository, but it doesn't insert anything > into > >>>>>> the tables in a transaction. > >>>>>> > >>>>>> If we're only seeing this in PostgreSQL and a workaround with > putting > >>>>>> JobOperator.start() outside a transaction works, I would suggest > that's > >>>>>> okay for now. I do agree it needs to be fixed, but we might want > to look > >>>>>> at how we're handling transaction in JBeret as a whole. The RI, > not that > >>>>>> I want to model anything after it, uses it's own > >>>>>> TransactionManagerAdapter. It might make sense for JBeret to use a > >>>>>> TransactionManager rather than a UserTransaction. Or put the > ownness on > >>>>>> the SPI implementation of the BatchEnvironment to handle the > transactions. > >>>>> > >>>>> Are you saying that the application should work around this by > calling a different bean method that is marked NOT_SUPPORTED to facilitate > suspending the JTA transaction? > >>>> No I'm just saying they need to invoke the JobOperator.start() > outside a transaction. At least from my understand on the JIRA that seems > to workaround the issue. I will admit to not fully looking into this in > detail though ;) > >>> > >>> That would be silly :) > >> > >> Requiring a NOT_SUPPORTED method that is. It's pretty easy for JBeret > to isolate the transaction if it wanted to > >> > >> tx = TransactionManager.suspend() > >> TransactionManager.begin() > >> // write the record > >> TransactionManager.commit() > >> TransactionManager.resume(tx); > >> > > > > What will happened to suspended Transaction when you get Exception on > TransactionManager.commit() ? > > You put resume in a finally block. Just like RequiresNew effectively > does. > > -- > Jason T. Greene > WildFly Lead / JBoss EAP Platform Architect > JBoss, a division of Red Hat > > > _______________________________________________ > wildfly-dev mailing list > wildfly-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/wildfly-dev > _______________________________________________ wildfly-dev mailing listwildfly-dev@lists.jboss.orghttps://lists.jboss.org/mailman/listinfo/wildfly-dev -- James R. Perkins Red Hat JBoss Middleware

If we aren’t suspending the transaction though, then we also need consistent behavior of store type. The point of my puzzle post is that changing the store type should not alter the behavior of that code. Also note that while this behavior *might* be more intuitive, its also more work, and it might eventually be in contradiction with the spec. Note IBM’s answer to a related question: https://java.net/projects/jbatch/lists/public/archive/2014-02/message/2 On Feb 5, 2014, at 10:41 AM, Stuart Douglas <stuart.w.douglas(a)gmail.com&gt; wrote: > Yes. I don't think we should be suspending the transaction. > > Stuart > > > On Wed, Feb 5, 2014 at 6:31 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: > You're suggesting something more like > > TransactionManager.getTransaction().registerSynchronization(synchronization); > > Rather than the; > > TransactionManager.suspend(); > try { > doStuff(); > } finally { > TransactionManager.resume(); > } > > Am I understanding that correctly? > > > On 02/04/2014 10:32 PM, Stuart Douglas wrote: >> This all sounds like a very similar problem to what we already do with EJB timers. Timers are transactional, if you create or cancel a timer it does not take effect until the transaction commits. >> >> The way this is accomplished is two fold: >> - The data store is transactional (or semi-transactional really in the case of the file data store, as we did not develop a fully transactional file system just for this) >> - Timers are not actually started or cancelled until the afterComplete() synchronization runs. >> >> I think it would make sense for JBeret to basically do the same. I think it would be very surprising to the user if jobs they started in transactions that abort just proceed as normal. >> >> Stuart >> >> >> On Tue, Feb 4, 2014 at 11:53 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: >> >> On Feb 4, 2014, at 3:51 PM, Radoslaw Rodak <rodakr(a)gmx.ch&gt; wrote: >> >>> Hi >>> >>> >>> Am 04.02.2014 um 22:16 schrieb Jason Greene <jason.greene(a)redhat.com&gt;: >>> >>>> >>>> On Feb 4, 2014, at 3:13 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: >>>> >>>>> >>>>> On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: >>>>> >>>>>> >>>>>> On 02/04/2014 12:40 PM, Scott Marlow wrote: >>>>>>> On 02/04/2014 02:42 PM, James R. Perkins wrote: >>>>>>>> >>>>>>>> On 02/04/2014 08:16 AM, Jason Greene wrote: >>>>>>>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >>>>>>>>> >>>>>>>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>>>>>>>>> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >>>>>>>>>>> >>>>>>>>>> yes, we could implement it in wildfly-batch integration module. >>>>>>>>>>> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. >>>>>>>>>> The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: >>>>>>>>>> >>>>>>>>>> Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: >>>>>>>>>> >>>>>>>>>> Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. >>>>>>>>> There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn’t describe it, and the RI doesn’t do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door. >>>>>>>> I don't see in the spec where it requires any kind of transaction around >>>>>>>> a job repository. In fact the spec states "Note the implementation of >>>>>>>> the job repository is outside the scope of this specification.". >>>>>>>> >>>>>>>> The RI does have a JDBC repository, but it doesn't insert anything into >>>>>>>> the tables in a transaction. >>>>>>>> >>>>>>>> If we're only seeing this in PostgreSQL and a workaround with putting >>>>>>>> JobOperator.start() outside a transaction works, I would suggest that's >>>>>>>> okay for now. I do agree it needs to be fixed, but we might want to look >>>>>>>> at how we're handling transaction in JBeret as a whole. The RI, not that >>>>>>>> I want to model anything after it, uses it's own >>>>>>>> TransactionManagerAdapter. It might make sense for JBeret to use a >>>>>>>> TransactionManager rather than a UserTransaction. Or put the ownness on >>>>>>>> the SPI implementation of the BatchEnvironment to handle the transactions. >>>>>>> >>>>>>> Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction? >>>>>> No I'm just saying they need to invoke the JobOperator.start() outside a transaction. At least from my understand on the JIRA that seems to workaround the issue. I will admit to not fully looking into this in detail though ;) >>>>> >>>>> That would be silly :) >>>> >>>> Requiring a NOT_SUPPORTED method that is. It’s pretty easy for JBeret to isolate the transaction if it wanted to >>>> >>>> tx = TransactionManager.suspend() >>>> TransactionManager.begin() >>>> // write the record >>>> TransactionManager.commit() >>>> TransactionManager.resume(tx); >>>> >>> >>> What will happened to suspended Transaction when you get Exception on TransactionManager.commit() ? >> >> You put resume in a finally block. Just like RequiresNew effectively does. >> >> -- >> Jason T. Greene >> WildFly Lead / JBoss EAP Platform Architect >> JBoss, a division of Red Hat >> >> >> _______________________________________________ >> wildfly-dev mailing list >> wildfly-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/wildfly-dev >> >> >> >> _______________________________________________ >> wildfly-dev mailing list >> >> wildfly-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/wildfly-dev > > -- > James R. Perkins > Red Hat JBoss Middleware > > -- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

On 02/05/2014 08:44 AM, Jason Greene wrote: > If we aren’t suspending the transaction though, then we also need consistent behavior of store type. The point of my puzzle post is that changing the store type should not alter the behavior of that code. Would using the suspend/resume now and changing to synchronization later be too big of a behavioral change?

From what I can tell the RI does nothing here. I think the main issue is using our JDBC repository. We insert records in a transaction which is where the issue was initially seen. My gut tells me for now to use the suspend/resume. Though I'd really like to do some more

reading/research on transactions. The spec is extremely vague on anything EE related. > > Also note that while this behavior *might* be more intuitive, its also more work, and it might eventually be in contradiction with the spec. Note IBM’s answer to a related question: > https://java.net/projects/jbatch/lists/public/archive/2014-02/message/2 > > On Feb 5, 2014, at 10:41 AM, Stuart Douglas <stuart.w.douglas(a)gmail.com&gt; wrote: > >> Yes. I don't think we should be suspending the transaction. >> >> Stuart >> >> >> On Wed, Feb 5, 2014 at 6:31 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: >> You're suggesting something more like >> >> TransactionManager.getTransaction().registerSynchronization(synchronization); >> >> Rather than the; >> >> TransactionManager.suspend(); >> try { >> doStuff(); >> } finally { >> TransactionManager.resume(); >> } >> >> Am I understanding that correctly? >> >> >> On 02/04/2014 10:32 PM, Stuart Douglas wrote: >>> This all sounds like a very similar problem to what we already do with EJB timers. Timers are transactional, if you create or cancel a timer it does not take effect until the transaction commits. >>> >>> The way this is accomplished is two fold: >>> - The data store is transactional (or semi-transactional really in the case of the file data store, as we did not develop a fully transactional file system just for this) >>> - Timers are not actually started or cancelled until the afterComplete() synchronization runs. >>> >>> I think it would make sense for JBeret to basically do the same. I think it would be very surprising to the user if jobs they started in transactions that abort just proceed as normal. >>> >>> Stuart >>> >>> >>> On Tue, Feb 4, 2014 at 11:53 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: >>> >>> On Feb 4, 2014, at 3:51 PM, Radoslaw Rodak <rodakr(a)gmx.ch&gt; wrote: >>> >>>> Hi >>>> >>>> >>>> Am 04.02.2014 um 22:16 schrieb Jason Greene <jason.greene(a)redhat.com&gt;: >>>> >>>>> On Feb 4, 2014, at 3:13 PM, Jason Greene <jason.greene(a)redhat.com&gt; wrote: >>>>> >>>>>> On Feb 4, 2014, at 3:01 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: >>>>>> >>>>>>> On 02/04/2014 12:40 PM, Scott Marlow wrote: >>>>>>>> On 02/04/2014 02:42 PM, James R. Perkins wrote: >>>>>>>>> On 02/04/2014 08:16 AM, Jason Greene wrote: >>>>>>>>>> On Feb 4, 2014, at 9:56 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: >>>>>>>>>> >>>>>>>>>>> On 2/4/14, 9:57 AM, Stuart Douglas wrote: >>>>>>>>>>>> I would use a transaction synchronization, so you don't spawn the other thread until the transaction is successfully committed. >>>>>>>>>>>> >>>>>>>>>>> yes, we could implement it in wildfly-batch integration module. >>>>>>>>>>>> What does the spec say about transactions? If a job is create in a thread that is part of a transaction and the transaction is rolled back should the job actually go ahead? Common sense would suggest not. >>>>>>>>>>> The transaction treatment in the batch spec is mostly around item processing, not much on how it interacts with the transaction in the running environment. The only place that it touches on Java EE environment is section 9.7 Transactionality: >>>>>>>>>>> >>>>>>>>>>> Chunk type check points are transactional. The batch runtime uses global transaction mode on the Java EE platform and local transaction mode on the Java SE platform. Global transaction timeout is configurable at step-level with a step-level property: >>>>>>>>>>> >>>>>>>>>>> Yes, I agree if the batch client side transaction is rolled back, the job execution should not proceed. With the current jberet impl, the job execution in this case will fail since the job repository is not in good state, like in the above bug. If we have transaction syncrhonization in place, then the job will not start running till transaction 1 is committed. >>>>>>>>>> There is a consistency problem here though. If you expect the client side to rollback on transaction failure, then the in-memory job store should as well. IMO before committing to such a big feature, I would recommend looking at what the RI does here. If the spec doesn’t describe it, and the RI doesn’t do it, then we should avoid investing time on it at least right now where we really need to get WF8 out the door. >>>>>>>>> I don't see in the spec where it requires any kind of transaction around >>>>>>>>> a job repository. In fact the spec states "Note the implementation of >>>>>>>>> the job repository is outside the scope of this specification.". >>>>>>>>> >>>>>>>>> The RI does have a JDBC repository, but it doesn't insert anything into >>>>>>>>> the tables in a transaction. >>>>>>>>> >>>>>>>>> If we're only seeing this in PostgreSQL and a workaround with putting >>>>>>>>> JobOperator.start() outside a transaction works, I would suggest that's >>>>>>>>> okay for now. I do agree it needs to be fixed, but we might want to look >>>>>>>>> at how we're handling transaction in JBeret as a whole. The RI, not that >>>>>>>>> I want to model anything after it, uses it's own >>>>>>>>> TransactionManagerAdapter. It might make sense for JBeret to use a >>>>>>>>> TransactionManager rather than a UserTransaction. Or put the ownness on >>>>>>>>> the SPI implementation of the BatchEnvironment to handle the transactions. >>>>>>>> Are you saying that the application should work around this by calling a different bean method that is marked NOT_SUPPORTED to facilitate suspending the JTA transaction? >>>>>>> No I'm just saying they need to invoke the JobOperator.start() outside a transaction. At least from my understand on the JIRA that seems to workaround the issue. I will admit to not fully looking into this in detail though ;) >>>>>> That would be silly :) >>>>> Requiring a NOT_SUPPORTED method that is. It’s pretty easy for JBeret to isolate the transaction if it wanted to >>>>> >>>>> tx = TransactionManager.suspend() >>>>> TransactionManager.begin() >>>>> // write the record >>>>> TransactionManager.commit() >>>>> TransactionManager.resume(tx); >>>>> >>>> What will happened to suspended Transaction when you get Exception on TransactionManager.commit() ? >>> You put resume in a finally block. Just like RequiresNew effectively does. >>> >>> -- >>> Jason T. Greene >>> WildFly Lead / JBoss EAP Platform Architect >>> JBoss, a division of Red Hat >>> >>> >>> _______________________________________________ >>> wildfly-dev mailing list >>> wildfly-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/wildfly-dev >>> >>> >>> >>> _______________________________________________ >>> wildfly-dev mailing list >>> >>> wildfly-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/wildfly-dev >> -- >> James R. Perkins >> Red Hat JBoss Middleware >> >> > -- > Jason T. Greene > WildFly Lead / JBoss EAP Platform Architect > JBoss, a division of Red Hat >

On Feb 4, 2014, at 1:42 PM, James R. Perkins <jperkins(a)redhat.com&gt; wrote: > I don't see in the spec where it requires any kind of transaction around a job repository. In fact the spec states "Note the implementation of the job repository is outside the scope of this specification.". > > The RI does have a JDBC repository, but it doesn't insert anything into the tables in a transaction. > > If we're only seeing this in PostgreSQL and a workaround with putting JobOperator.start() outside a transaction works, I would suggest that's okay for now. I do agree it needs to be fixed, but we might want to look at how we're handling transaction in JBeret as a whole. The RI, not that I want to model anything after it, uses it's own TransactionManagerAdapter. It might make sense for JBeret to use a TransactionManager rather than a UserTransaction. Or put the ownness on the SPI implementation of the BatchEnvironment to handle the transactions. Good points; and this brings me to my pathological batch puzzle. What does this code do: // Default gets transaction required String doStuff() { long execID = jobOperator.start("simplejob", props); BatchStatus status = jobOperator.getJobExecution(execID).getBatchStatus() while (status != COMMITTED && status != FAILED && status != ABANDONED) { Thread.sleep(1000); } return status.toString(); }

-- Jason T. Greene WildFly Lead / JBoss EAP Platform Architect JBoss, a division of Red Hat

A jberet user reported JBERET-29 <https://issues.jboss.org/browse/JBERET-29> (Foreign key constraint step_execution_jobexecutionid_fkey fails when using Postgresql on WildFly, and we are trying to fix it by jberet 1.0.0.Final. The problem happens when user app starts a job within a transaction (e.g., CMT EJB), jberet inserts JobExecution into database (thread 1 & transaction 1), and then spawn a jberet-batch thread to run the job (thead 2 & transaction 2). Sometimes T2 tries to access db before T1 is committed, hence the error reported by the user. What's the common approach for solving this kind of problem? I suppose other WildFly components may also have this issue and probably already solved. Using transaction synchronization is a cleaner solution than polling db, but I'm not sure about its full implication. Ideally, I don't want to use system level JTA API like TransactionManager or Synchronization in jberet proper, but probably we can implement it in WildFly jberet integration.

Thanks, Cheng _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

A jberet user reported JBERET-29 (Foreign key constraint step_execution_jobexecutionid_fkey fails when using Postgresql on WildFly, and we are trying to fix it by jberet 1.0.0.Final. The problem happens when user app starts a job within a transaction (e.g., CMT EJB), jberet inserts JobExecution into database (thread 1 & transaction 1), and then spawn a jberet-batch thread to run the job (thead 2 & transaction 2). Sometimes T2 tries to access db before T1 is committed, hence the error reported by the user. What's the common approach for solving this kind of problem? I suppose other WildFly components may also have this issue and probably already solved. Using transaction synchronization is a cleaner solution than polling db, but I'm not sure about its full implication. Ideally, I don't want to use system level JTA API like TransactionManager or Synchronization in jberet proper, but probably we can implement it in WildFly jberet integration. Thanks, Cheng _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev

On 2/4/14, 10:11 AM, Jason Greene wrote: > Does the spec require that JobExecution be tied to a CMT transaction? No. > It sounds like it should not be. If you were to make this into a local transaction that is committed before the call to create the job completes, then you shouldn’t have that problem. Are you saying with local transaction, it can access uncommitted data from another transaction? In this case, JobExecution is still not committed when it's accessed via foreign key by transaction 2.

Inside WildFly, we just look up and use UserTransaction to demarcate all transactions, and it seems a bit odd to use local transaction just for this case. The batch spec does require the use of global transaction on Java EE platform.

Cheng > On Feb 4, 2014, at 8:49 AM, Cheng Fang <cfang(a)redhat.com&gt; wrote: > >> A jberet user reported JBERET-29 (Foreign key constraint step_execution_jobexecutionid_fkey fails when using Postgresql on WildFly, and we are trying to fix it by jberet 1.0.0.Final. The problem happens when user app starts a job within a transaction (e.g., CMT EJB), jberet inserts JobExecution into database (thread 1 & transaction 1), and then spawn a jberet-batch thread to run the job (thead 2 & transaction 2). Sometimes T2 tries to access db before T1 is committed, hence the error reported by the user. >> >> What's the common approach for solving this kind of problem? I suppose other WildFly components may also have this issue and probably already solved. Using transaction synchronization is a cleaner solution than polling db, but I'm not sure about its full implication. Ideally, I don't want to use system level JTA API like TransactionManager or Synchronization in jberet proper, but probably we can implement it in WildFly jberet integration. >> >> Thanks, >> Cheng >> _______________________________________________ >> wildfly-dev mailing list >> wildfly-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/wildfly-dev > -- > Jason T. Greene > WildFly Lead / JBoss EAP Platform Architect > JBoss, a division of Red Hat > _______________________________________________ wildfly-dev mailing list wildfly-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/wildfly-dev