Created attachment 922506 [details] localhost.2014-07-30.log There is a regression in CR2 hibernate-dist.zip. The c3p0 connection pool is not working properly, causing NPEs in tomcat. (see attached log and EmployeeManager.java, HibernateUtil.java). I believe this to be caused by the new build process described in bug#1117235, that took place in between ER4 and CR1. (CR1 was not really tested, as it was not accepted by QE). Therefore, CR2 is the first build with the update c3p0 jar that we tested.
Created attachment 922507 [details] EmployeeManager.java
Created attachment 922508 [details] HibernateUtil.java
Regarding c3p0, I noticed that since we don't ship log4j with EAP, we never enabled log4j support. I have enabled it for this build. Also, we have been applying a patch (from Fedora) to add JDBC4 stubs and allow building against JDK 1.6, but this should not affect runtime. Please try again with the latest build <http://download.eng.bos.redhat.com/brewroot/repos/jb-eap-6-rhel-6-build/latest/maven/c3p0/c3p0/0.9.1.2-redhat-2/c3p0-0.9.1.2-redhat-2.jar>.
Created attachment 922877 [details] localhost.2014-07-31.log (In reply to David Walluck from comment #10) > Please try again with the latest build > <http://download.eng.bos.redhat.com/brewroot/repos/jb-eap-6-rhel-6-build/ > latest/maven/c3p0/c3p0/0.9.1.2-redhat-2/c3p0-0.9.1.2-redhat-2.jar>. The same results, unfortunately :-( But sometimes, are are some more exceptions in localhost.<date>.log
After discussing with Libor and Michal, currently what we could do is to revert hibernate back to the ones used in ER4. But that doesn't mean we should close this bug, and we have to fix this bug post 2.1.0 GA. In addition, reverting hibernate to the one in er4 will make bz1117235 back to known issue. brew untag-pkg dist-5E-ep-ews-2-candidate hibernate4-eap6-4.2.14-6.SP1_redhat_4.1.ep6.el5 hibernate4-eap6-4.2.14-5.SP1_redhat_3.1.ep6.el5 hibernate4-eap6-4.2.14-4.SP1_redhat_2.1.ep6.el5 brew untag-pkg jb-ews-2-rhel-6-candidate hibernate4-eap6-4.2.14-6.SP1_redhat_4.1.ep6.el6 hibernate4-eap6-4.2.14-5.SP1_redhat_3.1.ep6.el6 hibernate4-eap6-4.2.14-4.SP1_redhat_2.1.ep6.el6 brew untag-pkg jb-ews-2-rhel-7-candiate hibernate4-eap6-4.2.14-6.SP1_redhat_4.1.ep6.el7 hibernate4-eap6-4.2.14-5.SP1_redhat_3.1.ep6.el7 hibernate4-eap6-4.2.14-4.SP1_redhat_2.1.ep6.el7 Here are the hibernate in ER4: hibernate4-eap6-4.2.14-3.SP1_redhat_1.1.ep6.el5 hibernate4-eap6-4.2.14-3.SP1_redhat_1.1.ep6.el6 hibernate4-eap6-4.2.14-3.SP1_redhat_1.1.ep6.el7
hibernate4 reverted for zip: brew untag-pkg jb-eap-6-rhel-6-candidate hibernate4-4.2.14-3.SP1_redhat_2.1.ep6.el6 hibernate4-4.2.14-4.SP1_redhat_3.1.ep6.el6 hibernate4-4.2.14-5.SP1_redhat_4.1.ep6.el6 Reverted to this: hibernate4-4.2.14-2.SP1_redhat_1.1.ep6.el6 jb-eap-6-rhel-6-candidate mead-scheduler Michal, could you please help to pretest the zip is the correct version we should use: http://download.devel.redhat.com/brewroot/packages/hibernate4/4.2.14/2.SP1_redhat_1.1.ep6.el6/noarch/hibernate4-zip-4.2.14-2.SP1_redhat_1.1.ep6.el6.noarch.rpm This is the command to extract the zip: rpm2cpio hibernate4-zip-4.2.14-2.SP1_redhat_1.1.ep6.el6.noarch.rpm | cpio -idv Please note this zip will be cleaned up during jboss-ews process, so merely testing c3p0 part would be fine.
md5sums of two c3p0 jars are the same between ER4 hibernate and build from last comment
tested on MS Windows and all EWS Hibernate test passed
Let's keep this bug open until it's fixed. Move target release to TBD
This should be also fixed in JWS 3.0.0 CR3.1. It needs final verification.
VERIFIED with JWS 3.0.0 CR3.1