You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by "Rajith Attapattu (JIRA)" <qp...@incubator.apache.org> on 2007/05/02 18:52:15 UTC

[jira] Commented: (QPID-475) First step toward dtx support.

    [ https://issues.apache.org/jira/browse/QPID-475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12493142 ] 

Rajith Attapattu commented on QPID-475:
---------------------------------------

I have comitted the patch at rev 534541 on trunk.
There is one test case failing and I will try to fix it.
If not Arnuad will provide a patch soon.

Rajith

> First step toward dtx support. 
> -------------------------------
>
>                 Key: QPID-475
>                 URL: https://issues.apache.org/jira/browse/QPID-475
>             Project: Qpid
>          Issue Type: New Feature
>          Components: Java Broker
>    Affects Versions: M3
>            Reporter: Arnaud Simon
>         Assigned To: Rajith Attapattu
>             Fix For: M3
>
>         Attachments: BDB.patch, brokerpatch.patch, main-bdb.zip, main.zip
>
>
> This patch is a first step toward dtx support. 
> It includes the new Message store and transaction manager interface and a highly efficient BDB implementation of the store. The transaction manager implementation is also based on BDB. 
> This work must be extended for plugging a dtx classes that will simply interact with the transaction manager. 
> Note that I have tried not to change the existing architecture. I do think that we should carefully plan 0.10 support that may not be based on the existing architecture. 
> The configuration file must be changed to include the following store and transaction manager configuration tags: 
>    <store>
>                     <!-- <class>org.apache.qpid.server.store.berkeleydb.BDBMessageStore</class> -->
>                     <environment-path>${work}/localhost-store</environment-path> 
>                     <class>org.apache.qpid.server.store.berkeleydb.messageStore.MessageStoreImpl</class>
>                 </store>             
>                 <txn>
>                     <environment-tx-timeout>60</environment-tx-timeout>
>                 	  <class>org.apache.qpid.server.store.berkeleydb.txn.TransactionManagerImpl</class>
>                 </txn>
>               

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


Re: [java] test cases

Posted by Marnie McCormack <ma...@googlemail.com>.
Hi All,

Just to mention that there are already one/two JIRAs created for
writing/plugging in a persistent store and perhaps also for the JDBC/SQL
route.

Probably have low-ish numbers !

Great to see this stuff moving forward - thanks Arnauld !

Bfn,
Marnie
On 5/10/07, John O'Hara <jo...@gmail.com> wrote:
>
> You make it sound so easy </awe>
>
> john
>
> On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
> >
> > Hi Rob,
> >
> > I was thinking about that actually. I think I'll do it as this is a very
> > useful feature to have. So, I have added this task on my TODO list.
> >
> > Cheers
> >
> > Arnaud
> >
> >
> >
> >
> > On Thu, 2007-05-03 at 11:52 +0100, Robert Godfrey wrote:
> > > Hi Arnuad,
> > >
> > > what other stores are you writing?
> > >
> > > We *really* could do with a persistent store that has Apache
> > > compatible licensing...  even if it's not the most efficient it would
> > > at least enable us to add persistence testing to our unit tests.
> > >
> > > I started on looking at writing an JDBC/SQL based store for this
> > > purpose but would need to change it now you've made your changes.  If
> > > you're already looking at this, then I won't bother.
> > >
> > > Cheers,
> > > Rob
> > >
> > > On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
> > > > Hi,
> > > >
> > > > There was one path mistake in the config file that I checked in but
> > > > Gordon has spotted it and solved it (Thanks for that Gordon).
> > > >
> > > > Apart from that I am still working on having all the test cases
> > passing
> > > > against all the stores. So, I would encourage you not to use the
> > current
> > > > revision for production. I'll let you know when everything has been
> > > > tested. I am also adding crash recovery tests that are useful for
> > > > testing tx.
> > > >
> > > > I have also updated the in-memory store and tx manager that were not
> > > > properly implemented but required (I'll produce a new patch soon).
> > Note
> > > > that the BDB implementation is checked in under etp.108.redhat.com.
> I
> > > > have also other stores that are not yet finalized.
> > > >
> > > > One main thing that we need to change though is the way message
> chunks
> > > > are handled. At the minute this not really optimum.
> > > >
> > > > So, to summarize my plan for action for this week and next week is:
> > > > 1) finalize in memory store and tx
> > > > 2) have all the tests rewritten for the new store and tx interfaces
> > > > 3) have all the tests (Junit and system) passing against all the
> > stores
> > > > 4) add crash recovery tests
> > > > 5) extend dtx support
> > > > 6) finalize new stores
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> >
> >
>

Re: [java] test cases

Posted by John O'Hara <jo...@gmail.com>.
You make it sound so easy </awe>

john

On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
>
> Hi Rob,
>
> I was thinking about that actually. I think I'll do it as this is a very
> useful feature to have. So, I have added this task on my TODO list.
>
> Cheers
>
> Arnaud
>
>
>
>
> On Thu, 2007-05-03 at 11:52 +0100, Robert Godfrey wrote:
> > Hi Arnuad,
> >
> > what other stores are you writing?
> >
> > We *really* could do with a persistent store that has Apache
> > compatible licensing...  even if it's not the most efficient it would
> > at least enable us to add persistence testing to our unit tests.
> >
> > I started on looking at writing an JDBC/SQL based store for this
> > purpose but would need to change it now you've made your changes.  If
> > you're already looking at this, then I won't bother.
> >
> > Cheers,
> > Rob
> >
> > On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
> > > Hi,
> > >
> > > There was one path mistake in the config file that I checked in but
> > > Gordon has spotted it and solved it (Thanks for that Gordon).
> > >
> > > Apart from that I am still working on having all the test cases
> passing
> > > against all the stores. So, I would encourage you not to use the
> current
> > > revision for production. I'll let you know when everything has been
> > > tested. I am also adding crash recovery tests that are useful for
> > > testing tx.
> > >
> > > I have also updated the in-memory store and tx manager that were not
> > > properly implemented but required (I'll produce a new patch soon).
> Note
> > > that the BDB implementation is checked in under etp.108.redhat.com. I
> > > have also other stores that are not yet finalized.
> > >
> > > One main thing that we need to change though is the way message chunks
> > > are handled. At the minute this not really optimum.
> > >
> > > So, to summarize my plan for action for this week and next week is:
> > > 1) finalize in memory store and tx
> > > 2) have all the tests rewritten for the new store and tx interfaces
> > > 3) have all the tests (Junit and system) passing against all the
> stores
> > > 4) add crash recovery tests
> > > 5) extend dtx support
> > > 6) finalize new stores
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
>
>

Re: [java] test cases

Posted by Arnaud Simon <as...@redhat.com>.
Hi Rob,

I was thinking about that actually. I think I'll do it as this is a very
useful feature to have. So, I have added this task on my TODO list. 

Cheers

Arnaud




On Thu, 2007-05-03 at 11:52 +0100, Robert Godfrey wrote:
> Hi Arnuad,
> 
> what other stores are you writing?
> 
> We *really* could do with a persistent store that has Apache
> compatible licensing...  even if it's not the most efficient it would
> at least enable us to add persistence testing to our unit tests.
> 
> I started on looking at writing an JDBC/SQL based store for this
> purpose but would need to change it now you've made your changes.  If
> you're already looking at this, then I won't bother.
> 
> Cheers,
> Rob
> 
> On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
> > Hi,
> >
> > There was one path mistake in the config file that I checked in but
> > Gordon has spotted it and solved it (Thanks for that Gordon).
> >
> > Apart from that I am still working on having all the test cases passing
> > against all the stores. So, I would encourage you not to use the current
> > revision for production. I'll let you know when everything has been
> > tested. I am also adding crash recovery tests that are useful for
> > testing tx.
> >
> > I have also updated the in-memory store and tx manager that were not
> > properly implemented but required (I'll produce a new patch soon). Note
> > that the BDB implementation is checked in under etp.108.redhat.com. I
> > have also other stores that are not yet finalized.
> >
> > One main thing that we need to change though is the way message chunks
> > are handled. At the minute this not really optimum.
> >
> > So, to summarize my plan for action for this week and next week is:
> > 1) finalize in memory store and tx
> > 2) have all the tests rewritten for the new store and tx interfaces
> > 3) have all the tests (Junit and system) passing against all the stores
> > 4) add crash recovery tests
> > 5) extend dtx support
> > 6) finalize new stores
> >
> >
> >
> >
> >
> >
> >
> >


Re: [java] test cases

Posted by Robert Godfrey <ro...@gmail.com>.
Hi Arnuad,

what other stores are you writing?

We *really* could do with a persistent store that has Apache
compatible licensing...  even if it's not the most efficient it would
at least enable us to add persistence testing to our unit tests.

I started on looking at writing an JDBC/SQL based store for this
purpose but would need to change it now you've made your changes.  If
you're already looking at this, then I won't bother.

Cheers,
Rob

On 03/05/07, Arnaud Simon <as...@redhat.com> wrote:
> Hi,
>
> There was one path mistake in the config file that I checked in but
> Gordon has spotted it and solved it (Thanks for that Gordon).
>
> Apart from that I am still working on having all the test cases passing
> against all the stores. So, I would encourage you not to use the current
> revision for production. I'll let you know when everything has been
> tested. I am also adding crash recovery tests that are useful for
> testing tx.
>
> I have also updated the in-memory store and tx manager that were not
> properly implemented but required (I'll produce a new patch soon). Note
> that the BDB implementation is checked in under etp.108.redhat.com. I
> have also other stores that are not yet finalized.
>
> One main thing that we need to change though is the way message chunks
> are handled. At the minute this not really optimum.
>
> So, to summarize my plan for action for this week and next week is:
> 1) finalize in memory store and tx
> 2) have all the tests rewritten for the new store and tx interfaces
> 3) have all the tests (Junit and system) passing against all the stores
> 4) add crash recovery tests
> 5) extend dtx support
> 6) finalize new stores
>
>
>
>
>
>
>
>

[java] test cases

Posted by Arnaud Simon <as...@redhat.com>.
Hi,

There was one path mistake in the config file that I checked in but
Gordon has spotted it and solved it (Thanks for that Gordon). 

Apart from that I am still working on having all the test cases passing
against all the stores. So, I would encourage you not to use the current
revision for production. I'll let you know when everything has been
tested. I am also adding crash recovery tests that are useful for
testing tx. 

I have also updated the in-memory store and tx manager that were not
properly implemented but required (I'll produce a new patch soon). Note
that the BDB implementation is checked in under etp.108.redhat.com. I
have also other stores that are not yet finalized. 

One main thing that we need to change though is the way message chunks
are handled. At the minute this not really optimum.

So, to summarize my plan for action for this week and next week is: 
1) finalize in memory store and tx 
2) have all the tests rewritten for the new store and tx interfaces 
3) have all the tests (Junit and system) passing against all the stores 
4) add crash recovery tests 
5) extend dtx support 
6) finalize new stores