Hi All, http://lists.jboss.org/pipermail/wildfly-dev/2013-August/000574.html (thanks to Jason, Bill and Scott Marlow for the discussion) Jason has provided some objections to the JPA usage in PicketLink that can affect Wildfly boot time performance. He had raised this concern last time for the JPAIdentityStore implementation in PicketLink v2.5 which we have been trying to mitigate via the implementation of a JDBC driven Identity Store. The discussion happening in the thread above however is for the JPABasedTokenRegistry and JPABasedRevocationRegistry part of v2.1.x which predates PicketLink v2.5 https://github.com/picketlink2/federation/blob/master/picketlink-core/src... https://github.com/picketlink2/federation/blob/master/picketlink-core/src... I think we should also implement a straight JDBC based token registry and revocation registry to avoid the Wildfly boot drag. Regards, Anil

Anil, I surely can, will have two POC environments (PL cluster + MySQL cluster) for many weeks. Thank you very much for the attention to the issue, I really appreciated it. Regards, On Tue, Aug 6, 2013 at 12:54 PM, Anil Saldhana <Anil.Saldhana(a)redhat.com <mailto:Anil.Saldhana@redhat.com>> wrote: https://issues.jboss.org/browse/PLINK2-101 https://issues.jboss.org/browse/PLINK2-102 I hope Fernando can help test this. :) On 08/06/2013 10:30 AM, Anil Saldhana wrote: Hi All, http://lists.jboss.org/pipermail/wildfly-dev/2013-August/000574.html (thanks to Jason, Bill and Scott Marlow for the discussion) Jason has provided some objections to the JPA usage in PicketLink that can affect Wildfly boot time performance. He had raised this concern last time for the JPAIdentityStore implementation in PicketLink v2.5 which we have been trying to mitigate via the implementation of a JDBC driven Identity Store. The discussion happening in the thread above however is for the JPABasedTokenRegistry and JPABasedRevocationRegistry part of v2.1.x which predates PicketLink v2.5 https://github.com/picketlink2/federation/blob/master/picketlink-core/src... https://github.com/picketlink2/federation/blob/master/picketlink-core/src... I think we should also implement a straight JDBC based token registry and revocation registry to avoid the Wildfly boot drag. Regards, Anil

Great, I have updated PLINK2-97 and asked the WF team what to do with WFLY-1691. Regards. On Tue, Aug 6, 2013 at 6:11 PM, Anil Saldhana <Anil.Saldhana(a)redhat.com <mailto:Anil.Saldhana@redhat.com>> wrote: Fernando, finished the code. I am going to test with a DB on my laptop before checking. Regards, Anil On 08/06/2013 03:52 PM, Fernando Ribeiro wrote: > Anil, > > I surely can, will have two POC environments (PL cluster + MySQL > cluster) for many weeks. > > Thank you very much for the attention to the issue, I really > appreciated it. > > Regards, > > > On Tue, Aug 6, 2013 at 12:54 PM, Anil Saldhana > <Anil.Saldhana(a)redhat.com <mailto:Anil.Saldhana@redhat.com>> wrote: > > https://issues.jboss.org/browse/PLINK2-101 > https://issues.jboss.org/browse/PLINK2-102 > > I hope Fernando can help test this. :) > > On 08/06/2013 10:30 AM, Anil Saldhana wrote: > > Hi All, > http://lists.jboss.org/pipermail/wildfly-dev/2013-August/000574.html > (thanks to Jason, Bill and Scott Marlow for the discussion) > > Jason has provided some objections to the JPA usage in > PicketLink that > can affect Wildfly boot time performance. He had raised > this concern > last time for the JPAIdentityStore implementation in > PicketLink v2.5 > which we have been trying to mitigate via the > implementation of a JDBC > driven Identity Store. > > The discussion happening in the thread above however is > for the > JPABasedTokenRegistry and JPABasedRevocationRegistry part > of v2.1.x > which predates PicketLink v2.5 > https://github.com/picketlink2/federation/blob/master/picketlink-core/src... > https://github.com/picketlink2/federation/blob/master/picketlink-core/src... > > I think we should also implement a straight JDBC based > token registry > and revocation registry to avoid the Wildfly boot drag. > > Regards, > Anil > >