[JBoss-dev] jbossweb build.15 Build Successful
View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jbossweb?log=log20060323220025Lbuild.15 BUILD COMPLETE - build.15Date of build: 03/23/2006 22:00:25Time to build: 24 minutes 3 seconds Unit Tests: (0) Total Errors and Failures: (0) Modifications since last build: (first 50 of 0)
Re: [JBoss-dev] how do Class-Path entries effect deployment?
Can you take a look at the forum post? It looks like the Class-path entries are being scanned. Not sure how that is possible since di.url is being used as the scanned thing. Scott M Stark wrote: The referenced jars are added to the DeplomentInfo's class loader classapth. Any other standard class loaders such as URLClassLoaders that are created for the jar will also pickup these as part of their classpath. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bill Burke Sent: Thursday, March 23, 2006 2:39 PM To: Jboss-Dev Subject: [JBoss-dev] how do Class-Path entries effect deployment? User running into severe deployment problems. http://www.jboss.com/index.html?module=bb&op=viewtopic&p=39322 57#3932257 -- Bill Burke Chief Architect JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&; dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development -- Bill Burke Chief Architect JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: jboss-4.0-testsuite Build Completed With Testsuite Errors
>>I'm not seeing the jbossretro-rt.jar anywhere in the current 4.0 build, and my workspace is up to date. Has this been integrated into the dist structure yet? http://fisheye.jboss.com/changelog/JBoss/build/jboss?cs=Branch_4_0:rcampbell:20060323050136 I see it in the cruisecontrol checkout: /services/cruisecontrol/work/checkout/jboss-4.0-testsuite/build/output/jboss-4.0.4.CR2/server/all/lib [EMAIL PROTECTED] lib]$ ls jbossretro-rt.jar jbossretro-rt.jar From: Scott M Stark Sent: Thursday, March 23, 2006 7:31 PM To: Scott M Stark; Ryan Campbell; '[EMAIL PROTECTED]'; 'jboss-development@lists.sourceforge.net' Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors This is failing under 1.4.2 because javassist is not getting pulled into the jacc config: 2006-03-23 15:59:02,535 ERROR [org.jboss.deployment.MainDeployer] Could not create deployment: file:/services/cruisecontrol/checkout/jboss-4.0-testsuite/testsuite/output/lib/cmp2-audit.jar java.lang.NoClassDefFoundError: javassist/NotFoundException at org.jboss.ws.server.WebServiceDeployerEJB21.isWebserviceDeployment(WebServiceDeployerEJB21.java:76) at org.jboss.ws.server.WebServiceDeployer.create(WebServiceDeployer.java:101) javassist should be integrated via the server lib dir rather than the aop deployer due to the jbossretro, hibernate, etc dependencies not related to aop. I'm not seeing the jbossretro-rt.jar anywhere in the current 4.0 build, and my workspace is up to date. Has this been integrated into the dist structure yet? From: Scott M Stark Sent: Thursday, March 23, 2006 4:50 PM To: Ryan Campbell; '[EMAIL PROTECTED]'; 'jboss-development@lists.sourceforge.net' Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors That is an irrelevant debug level log message indicating that the does not have a setter for injecting the xmbean object name. It has no affect on the jacc service. I'm looking at the failures. From: Ryan Campbell Sent: Thursday, March 23, 2006 3:03 PM To: [EMAIL PROTECTED]; 'jboss-development@lists.sourceforge.net' Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors All these JACC tests are failing, and the jacc-security-mgr logs show this: java.lang.NoSuchMethodException: org.jboss.security.jacc.SecurityService.setObjectName(javax.management.ObjectName) at java.lang.Class.getMethod(Class.java:986) at org.jboss.mx.server.AbstractMBeanInvoker.inject(AbstractMBeanInvoker.java:944) at org.jboss.mx.server.AbstractMBeanInvoker.preRegister(AbstractMBeanInvoker.java:658) at org.jboss.mx.server.registry.BasicMBeanRegistry.invokePreRegister(BasicMBeanRegistry.java:697) at org.jboss.mx.server.registry.BasicMBeanRegistry.registerMBean(BasicMBeanRegistry.java:211) at sun.reflect.GeneratedMethodAccessor1.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:324) at org.jboss.mx.interceptor.ReflectedDispatcher.invoke(ReflectedDispatcher.java:155) at org.jboss.mx.server.Invocation.dispatch(Invocation.java:94) at org.jboss.mx.interceptor.AbstractInterceptor.invoke(AbstractInterceptor.java:133) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.interceptor.ModelMBeanOperationInterceptor.invoke(ModelMBeanOperationInterceptor.java:142) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.server.AbstractMBeanInvoker.invoke(AbstractMBeanInvoker.java:260) at org.jboss.mx.server.MBeanServerImpl.invoke(MBeanServerImpl.java:659) at org.jboss.mx.server.MBeanServerImpl$3.run(MBeanServerImpl.java:1422) at java.security.AccessController.doPrivileged(Native Method) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1417) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1350) at org.jboss.mx.server.MBeanServerImpl.createMBean(MBeanServerImpl.java:345) at org.jboss.system.ServiceCreator.install(ServiceCreator.java:181) From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Thursday, March 23, 2006 3:12 PM To: Adrian Brock; Anil Saldhana; Bill Burke; Brian Stansberry; jboss-development@lists.sourceforge.net; [EMAIL PROTECTED]; QA; Ryan Campbell; Ruel Loehr; Scott M Stark; Thomas Diesler Subject: jboss-4.0-testsuite Build Completed With Testsuite Errors Importance: High View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.0-testsuite?log=log20060323142936 TESTS FAILED Ant Error Message: /services/cruisecontrol/work/scripts/build-jboss-common.xml:235: The following error occurred while executing this line: /services/cruisecontrol/work/scripts/build-common-targets.xml:26: Build Successful - Tests completed w
[JBoss-dev] RE: jboss-4.0-testsuite Build Completed With Testsuite Errors
This is failing under 1.4.2 because javassist is not getting pulled into the jacc config: 2006-03-23 15:59:02,535 ERROR [org.jboss.deployment.MainDeployer] Could not create deployment: file:/services/cruisecontrol/checkout/jboss-4.0-testsuite/testsuite/output/lib/cmp2-audit.jarjava.lang.NoClassDefFoundError: javassist/NotFoundException at org.jboss.ws.server.WebServiceDeployerEJB21.isWebserviceDeployment(WebServiceDeployerEJB21.java:76) at org.jboss.ws.server.WebServiceDeployer.create(WebServiceDeployer.java:101) javassist should be integrated via the server lib dir rather than the aop deployer due to the jbossretro, hibernate, etc dependencies not related to aop. I'm not seeing the jbossretro-rt.jar anywhere in the current 4.0 build, and my workspace is up to date. Has this been integrated into the dist structure yet? From: Scott M Stark Sent: Thursday, March 23, 2006 4:50 PMTo: Ryan Campbell; '[EMAIL PROTECTED]'; 'jboss-development@lists.sourceforge.net'Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors That is an irrelevant debug level log message indicating that the does not have a setter for injecting the xmbean object name. It has no affect on the jacc service. I'm looking at the failures. From: Ryan Campbell Sent: Thursday, March 23, 2006 3:03 PMTo: [EMAIL PROTECTED]; 'jboss-development@lists.sourceforge.net'Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors All these JACC tests are failing, and the jacc-security-mgr logs show this: java.lang.NoSuchMethodException: org.jboss.security.jacc.SecurityService.setObjectName(javax.management.ObjectName) at java.lang.Class.getMethod(Class.java:986) at org.jboss.mx.server.AbstractMBeanInvoker.inject(AbstractMBeanInvoker.java:944) at org.jboss.mx.server.AbstractMBeanInvoker.preRegister(AbstractMBeanInvoker.java:658) at org.jboss.mx.server.registry.BasicMBeanRegistry.invokePreRegister(BasicMBeanRegistry.java:697) at org.jboss.mx.server.registry.BasicMBeanRegistry.registerMBean(BasicMBeanRegistry.java:211) at sun.reflect.GeneratedMethodAccessor1.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:324) at org.jboss.mx.interceptor.ReflectedDispatcher.invoke(ReflectedDispatcher.java:155) at org.jboss.mx.server.Invocation.dispatch(Invocation.java:94) at org.jboss.mx.interceptor.AbstractInterceptor.invoke(AbstractInterceptor.java:133) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.interceptor.ModelMBeanOperationInterceptor.invoke(ModelMBeanOperationInterceptor.java:142) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.server.AbstractMBeanInvoker.invoke(AbstractMBeanInvoker.java:260) at org.jboss.mx.server.MBeanServerImpl.invoke(MBeanServerImpl.java:659) at org.jboss.mx.server.MBeanServerImpl$3.run(MBeanServerImpl.java:1422) at java.security.AccessController.doPrivileged(Native Method) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1417) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1350) at org.jboss.mx.server.MBeanServerImpl.createMBean(MBeanServerImpl.java:345) at org.jboss.system.ServiceCreator.install(ServiceCreator.java:181) From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Thursday, March 23, 2006 3:12 PMTo: Adrian Brock; Anil Saldhana; Bill Burke; Brian Stansberry; jboss-development@lists.sourceforge.net; [EMAIL PROTECTED]; QA; Ryan Campbell; Ruel Loehr; Scott M Stark; Thomas DieslerSubject: jboss-4.0-testsuite Build Completed With Testsuite ErrorsImportance: High View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.0-testsuite?log=log20060323142936 TESTS FAILED Ant Error Message: /services/cruisecontrol/work/scripts/build-jboss-common.xml:235: The following error occurred while executing this line: /services/cruisecontrol/work/scripts/build-common-targets.xml:26: Build Successful - Tests completed with errors or failures. Date of build: 03/23/2006 14:29:36 Time to build: 98 minutes 20 seconds Last changed: 03/23/2006 14:22:36 Last log entry: fix uneeded dependency Unit
[JBoss-dev] RE: jboss-4.0-testsuite Build Completed With Testsuite Errors
That is an irrelevant debug level log message indicating that the does not have a setter for injecting the xmbean object name. It has no affect on the jacc service. I'm looking at the failures. From: Ryan Campbell Sent: Thursday, March 23, 2006 3:03 PMTo: [EMAIL PROTECTED]; 'jboss-development@lists.sourceforge.net'Subject: RE: jboss-4.0-testsuite Build Completed With Testsuite Errors All these JACC tests are failing, and the jacc-security-mgr logs show this: java.lang.NoSuchMethodException: org.jboss.security.jacc.SecurityService.setObjectName(javax.management.ObjectName) at java.lang.Class.getMethod(Class.java:986) at org.jboss.mx.server.AbstractMBeanInvoker.inject(AbstractMBeanInvoker.java:944) at org.jboss.mx.server.AbstractMBeanInvoker.preRegister(AbstractMBeanInvoker.java:658) at org.jboss.mx.server.registry.BasicMBeanRegistry.invokePreRegister(BasicMBeanRegistry.java:697) at org.jboss.mx.server.registry.BasicMBeanRegistry.registerMBean(BasicMBeanRegistry.java:211) at sun.reflect.GeneratedMethodAccessor1.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:324) at org.jboss.mx.interceptor.ReflectedDispatcher.invoke(ReflectedDispatcher.java:155) at org.jboss.mx.server.Invocation.dispatch(Invocation.java:94) at org.jboss.mx.interceptor.AbstractInterceptor.invoke(AbstractInterceptor.java:133) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.interceptor.ModelMBeanOperationInterceptor.invoke(ModelMBeanOperationInterceptor.java:142) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.server.AbstractMBeanInvoker.invoke(AbstractMBeanInvoker.java:260) at org.jboss.mx.server.MBeanServerImpl.invoke(MBeanServerImpl.java:659) at org.jboss.mx.server.MBeanServerImpl$3.run(MBeanServerImpl.java:1422) at java.security.AccessController.doPrivileged(Native Method) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1417) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1350) at org.jboss.mx.server.MBeanServerImpl.createMBean(MBeanServerImpl.java:345) at org.jboss.system.ServiceCreator.install(ServiceCreator.java:181) From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Thursday, March 23, 2006 3:12 PMTo: Adrian Brock; Anil Saldhana; Bill Burke; Brian Stansberry; jboss-development@lists.sourceforge.net; [EMAIL PROTECTED]; QA; Ryan Campbell; Ruel Loehr; Scott M Stark; Thomas DieslerSubject: jboss-4.0-testsuite Build Completed With Testsuite ErrorsImportance: High View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.0-testsuite?log=log20060323142936 TESTS FAILED Ant Error Message: /services/cruisecontrol/work/scripts/build-jboss-common.xml:235: The following error occurred while executing this line: /services/cruisecontrol/work/scripts/build-common-targets.xml:26: Build Successful - Tests completed with errors or failures. Date of build: 03/23/2006 14:29:36 Time to build: 98 minutes 20 seconds Last changed: 03/23/2006 14:22:36 Last log entry: fix uneeded dependency Unit Tests: (2879) Total Errors and Failures: (234)
RE: [JBoss-dev] how do Class-Path entries effect deployment?
The referenced jars are added to the DeplomentInfo's class loader classapth. Any other standard class loaders such as URLClassLoaders that are created for the jar will also pickup these as part of their classpath. > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On > Behalf Of Bill Burke > Sent: Thursday, March 23, 2006 2:39 PM > To: Jboss-Dev > Subject: [JBoss-dev] how do Class-Path entries effect deployment? > > User running into severe deployment problems. > > http://www.jboss.com/index.html?module=bb&op=viewtopic&p=39322 > 57#3932257 > > > -- > Bill Burke > Chief Architect > JBoss Inc. > > > --- > This SF.Net email is sponsored by xPML, a groundbreaking > scripting language > that extends applications into web and mobile media. Attend > the live webcast > and join the prime developer group breaking into this new > coding territory! > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&; > dat=121642 > ___ > JBoss-Development mailing list > JBoss-Development@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/jboss-development > --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] Re: jboss-4.0-testsuite Build Completed With Testsuite Errors
Hmmm, probably related to the EJB3 jacc failures? Ryan Campbell wrote: All these JACC tests are failing, and the jacc-security-mgr logs show this: java.lang.NoSuchMethodException: org.jboss.security.jacc.SecurityService.setObjectName(javax.management.ObjectName) at java.lang.Class.getMethod(Class.java:986) at org.jboss.mx.server.AbstractMBeanInvoker.inject(AbstractMBeanInvoker.java:944) at org.jboss.mx.server.AbstractMBeanInvoker.preRegister(AbstractMBeanInvoker.java:658) at org.jboss.mx.server.registry.BasicMBeanRegistry.invokePreRegister(BasicMBeanRegistry.java:697) at org.jboss.mx.server.registry.BasicMBeanRegistry.registerMBean(BasicMBeanRegistry.java:211) at sun.reflect.GeneratedMethodAccessor1.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:324) at org.jboss.mx.interceptor.ReflectedDispatcher.invoke(ReflectedDispatcher.java:155) at org.jboss.mx.server.Invocation.dispatch(Invocation.java:94) at org.jboss.mx.interceptor.AbstractInterceptor.invoke(AbstractInterceptor.java:133) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.interceptor.ModelMBeanOperationInterceptor.invoke(ModelMBeanOperationInterceptor.java:142) at org.jboss.mx.server.Invocation.invoke(Invocation.java:88) at org.jboss.mx.server.AbstractMBeanInvoker.invoke(AbstractMBeanInvoker.java:260) at org.jboss.mx.server.MBeanServerImpl.invoke(MBeanServerImpl.java:659) at org.jboss.mx.server.MBeanServerImpl$3.run(MBeanServerImpl.java:1422) at java.security.AccessController.doPrivileged(Native Method) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1417) at org.jboss.mx.server.MBeanServerImpl.registerMBean(MBeanServerImpl.java:1350) at org.jboss.mx.server.MBeanServerImpl.createMBean(MBeanServerImpl.java:345) at org.jboss.system.ServiceCreator.install(ServiceCreator.java:181) *From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *Sent:* Thursday, March 23, 2006 3:12 PM *To:* Adrian Brock; Anil Saldhana; Bill Burke; Brian Stansberry; jboss-development@lists.sourceforge.net; [EMAIL PROTECTED]; QA; Ryan Campbell; Ruel Loehr; Scott M Stark; Thomas Diesler *Subject:* jboss-4.0-testsuite Build Completed With Testsuite Errors *Importance:* High View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.0-testsuite?log=log20060323142936 *TESTS FAILED* *Ant Error Message: */services/cruisecontrol/work/scripts/build-jboss-common.xml:235: The following error occurred while executing this line: /services/cruisecontrol/work/scripts/build-common-targets.xml:26: Build Successful - Tests completed with errors or failures. *Date of build: *03/23/2006 14:29:36 *Time to build: *98 minutes 20 seconds *Last changed: *03/23/2006 14:22:36 *Last log entry: *fix uneeded dependency Unit Tests: (2879) Total Errors and Failures: (234) testServerFound org.jboss.test.cluster.test.UndeployTestCase(Default-UDP) testRelations org.jboss.test.cmp2.cmrstress.CMRStressTestCase(JACC+SecurityMgr) testRelations org.jboss.test.cmp2.cmrstress.CMRStressTestCase(JACC) testCMRTransaction org.jboss.test.cmp2.cmrtransaction.test.CMRTransactionUnitTestCase(JACC+SecurityMgr) testCMRTransaction org.jboss.test.cmp2.cmrtransaction.test.CMRTransactionUnitTestCase(JACC) testLazyResultSetLoading org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testCascadeDelete org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testCategory_Type org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testCategory_Type_BatchCascadeDelete org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testTxTester_none org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testJBossQL org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testEJBQL org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testCompiler org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testLimitOffset org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) testFinderWithLimitOffset org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) test_setInPostCreate org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) test_dvo org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) test_getOrdersShippedToCA org.jboss.test.cmp2.commerce.CompleteUnitTestCase(JACC+SecurityMgr) test_ge
[JBoss-dev] how do Class-Path entries effect deployment?
User running into severe deployment problems. http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3932257#3932257 -- Bill Burke Chief Architect JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
On Thu, 2006-03-23 at 09:52 -0600, Clebert Suconic wrote: > 1) It is slow > - It's the best I can do now, since there is no way on the API to get > such thing. I can verify later if there is a way to associate two tags > with a class on JVMTI, that would help me doing a faster navigation. > I'm not familiar with JVMTI and the doco is not very good, at least for "newbies" trying to understands the semantics of the call. I need to expermient. ;-) > 2) The formatting is not good > - Any ideas on how to improve it? I have just made a quick fix, > improved a little bit, but not much I'd like to see something like reference <- referer1.field <- referer2.field <- referer3.field And although using object.toString() is useful, it can be obscure when it doesn't include [EMAIL PROTECTED] like some of the jsr77 classes and it is totally uninformative for arrays. Similarly, some collections like linked lists or HashMaps don't show up well with the algorithm you are using. > > 3) Showing the field name > - I don't have a way on the event chain to determine what's the field > keeping a reference. It would be nice though. > http://java.sun.com/j2se/1.5.0/docs/guide/jvmti/jvmti.html#jvmtiObjectReferenceCallback http://java.sun.com/j2se/1.5.0/docs/guide/jvmti/jvmti.html#JVMTI_REFERENCE_FIELD -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
- My first implementation was an InMemory model. It was fast and beautiful, but I couldn't process large snapshots. - My second implementation was then HSQLDB. But when I was navigating, if I needed for instance to lookup on references for Strings, I had to look for reference holders on more than a thousand objects. The query was fast, but since I needed to execute it several times, took me a lot of time to process. Then I tried to use dynamic queries (if ID=1 or id=2 or id=3. it=100) and couldn't find a way to process - My third implementation (current) now is to use what I called FileCollections with a dummy quick sort. Since I was using just quickSort/binarySearch that was way faster than using HSQLDB. -Original Message- From: Adrian Brock Sent: Thursday, March 23, 2006 9:44 AM To: Clebert Suconic Cc: Scott M Stark; QA; jboss-development@lists.sourceforge.net Subject: RE: JBAS-2972 - OOME / Redeployment leakages On Thu, 2006-03-23 at 09:23 -0600, Clebert Suconic wrote: > > I can imagine a visitor pattern that understands what is > permenantly deployed and what is hotdeployed and determines > the root cause. > > I have that already. I can create a heapSnapshot, and analyze it later. > But at this point the method I have to sort the will takes a lot of time > to process now. I need a faster way to index file (without using a > database). > Can't you just use hsqldb in memory? That is what it was designed for ;-) -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
1) It is slow - It's the best I can do now, since there is no way on the API to get such thing. I can verify later if there is a way to associate two tags with a class on JVMTI, that would help me doing a faster navigation. 2) The formatting is not good - Any ideas on how to improve it? I have just made a quick fix, improved a little bit, but not much 3) Showing the field name - I don't have a way on the event chain to determine what's the field keeping a reference. It would be nice though. -Original Message- From: Adrian Brock Sent: Thursday, March 23, 2006 9:46 AM To: Clebert Suconic Cc: Scott M Stark; QA; jboss-development@lists.sourceforge.net Subject: RE: JBAS-2972 - OOME / Redeployment leakages On Thu, 2006-03-23 at 09:23 -0600, Clebert Suconic wrote: > The new features uses getRerenceHolders Method, which does an inverted > walkthrough everyObject looking for references on the object passed by > parameter, what can be done on a life JVM (I can discover holders > without create a snapshot) > > The method takes probably one second to run every time you call it, > since it walks through the whole heap. I have to do that, since there > isn't a way to fire reference events by its referenced objects. I have > to calculate the equivalent backward operation. For a 10 levels analysis > on a classLoader will probably take 1 minute or two. > I've been look at that this morning. 1) It is slow 2) The formatting is not good 3) Showing the field name There must be a better way to implement it, since that is what the api was designed for. > > BTW: If someone is using JVMTIInterface, I will create a xmdesc to the > MBean today/tomorrow. Right now you will have to look at parameter names > on source code (JVMTIInterfaceMBean) > > > > > > -Original Message- > From: Adrian Brock > Sent: Thursday, March 23, 2006 4:27 AM > To: Clebert Suconic > Cc: Scott M Stark; QA > Subject: RE: JBAS-2972 - OOME / Redeployment leakages > > On Thu, 2006-03-23 at 01:37 -0600, Clebert Suconic wrote: > > I created another method on JVMTIInterface, called > > exploreClassRefernces. With that method you can use JMXConsole to > > explore class and classLoader references to a class (given by name). > > > > - First let me explain how the data is organized: > > > > > > ReferencesToClassLoader.html shows every reference to > > [EMAIL PROTECTED] url=null > > ,addedOrder=50}, on this case the ClassLoader responsible for > deploying > > testbyvalue.ear. > > > > > > I have limited this analysis up to 10 levels. > > A better approach would be dump all objects and their > references out to a file. Then write something > that runs a full gc() over that information. > > Once that is done, you need something that lets you "walk the tree". > > I can imagine a visitor pattern that understands what is > permenantly deployed and what is hotdeployed and determines > the root cause. > > Otherwise, a simple GUI would do the trick. -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
On Thu, 2006-03-23 at 09:23 -0600, Clebert Suconic wrote: > The new features uses getRerenceHolders Method, which does an inverted > walkthrough everyObject looking for references on the object passed by > parameter, what can be done on a life JVM (I can discover holders > without create a snapshot) > > The method takes probably one second to run every time you call it, > since it walks through the whole heap. I have to do that, since there > isn't a way to fire reference events by its referenced objects. I have > to calculate the equivalent backward operation. For a 10 levels analysis > on a classLoader will probably take 1 minute or two. > I've been look at that this morning. 1) It is slow 2) The formatting is not good 3) Showing the field name There must be a better way to implement it, since that is what the api was designed for. > > BTW: If someone is using JVMTIInterface, I will create a xmdesc to the > MBean today/tomorrow. Right now you will have to look at parameter names > on source code (JVMTIInterfaceMBean) > > > > > > -Original Message- > From: Adrian Brock > Sent: Thursday, March 23, 2006 4:27 AM > To: Clebert Suconic > Cc: Scott M Stark; QA > Subject: RE: JBAS-2972 - OOME / Redeployment leakages > > On Thu, 2006-03-23 at 01:37 -0600, Clebert Suconic wrote: > > I created another method on JVMTIInterface, called > > exploreClassRefernces. With that method you can use JMXConsole to > > explore class and classLoader references to a class (given by name). > > > > - First let me explain how the data is organized: > > > > > > ReferencesToClassLoader.html shows every reference to > > [EMAIL PROTECTED] url=null > > ,addedOrder=50}, on this case the ClassLoader responsible for > deploying > > testbyvalue.ear. > > > > > > I have limited this analysis up to 10 levels. > > A better approach would be dump all objects and their > references out to a file. Then write something > that runs a full gc() over that information. > > Once that is done, you need something that lets you "walk the tree". > > I can imagine a visitor pattern that understands what is > permenantly deployed and what is hotdeployed and determines > the root cause. > > Otherwise, a simple GUI would do the trick. -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
On Thu, 2006-03-23 at 09:23 -0600, Clebert Suconic wrote: > > I can imagine a visitor pattern that understands what is > permenantly deployed and what is hotdeployed and determines > the root cause. > > I have that already. I can create a heapSnapshot, and analyze it later. > But at this point the method I have to sort the will takes a lot of time > to process now. I need a faster way to index file (without using a > database). > Can't you just use hsqldb in memory? That is what it was designed for ;-) -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: FW: JBAS-2972 - OOME / Redeployment leakages
On Thu, 2006-03-23 at 09:05 -0600, Clebert Suconic wrote: > > I don't even know what ThreadPool changes you are referring to, are > they real or proposed? > > I thought we had some changes on the ThreadPools. My bad. > I meant that because of references through ThreadWithAttributes (they > will be always appear for some reason on JVMTI). > > > Anyway, I need to know if a Thread is cleared > (Thread.setContextClassLoader(somethingElse) before the Thread goes to > sleep. > If the Thread keeps the ClassLoader reference until next request, that > will make my life harder creating a TestCase structure for testing > redeployment/classLeakages. All code should be switching and restoring the classloader during the invocations. Anything else is a bug. ClassLoader old = tcl; tcl = myClassLoader try { } finally { tcl = old; } Of course if something (e.g. an MBean) that is hotdeployed starts a thread and doesn't stop it at undeployment then it will leak the classloader. But that is also a bug. -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: JBAS-2972 - OOME / Redeployment leakages
> I can imagine a visitor pattern that understands what is permenantly deployed and what is hotdeployed and determines the root cause. I have that already. I can create a heapSnapshot, and analyze it later. But at this point the method I have to sort the will takes a lot of time to process now. I need a faster way to index file (without using a database). The new features uses getRerenceHolders Method, which does an inverted walkthrough everyObject looking for references on the object passed by parameter, what can be done on a life JVM (I can discover holders without create a snapshot) The method takes probably one second to run every time you call it, since it walks through the whole heap. I have to do that, since there isn't a way to fire reference events by its referenced objects. I have to calculate the equivalent backward operation. For a 10 levels analysis on a classLoader will probably take 1 minute or two. BTW: If someone is using JVMTIInterface, I will create a xmdesc to the MBean today/tomorrow. Right now you will have to look at parameter names on source code (JVMTIInterfaceMBean) -Original Message- From: Adrian Brock Sent: Thursday, March 23, 2006 4:27 AM To: Clebert Suconic Cc: Scott M Stark; QA Subject: RE: JBAS-2972 - OOME / Redeployment leakages On Thu, 2006-03-23 at 01:37 -0600, Clebert Suconic wrote: > I created another method on JVMTIInterface, called > exploreClassRefernces. With that method you can use JMXConsole to > explore class and classLoader references to a class (given by name). > > - First let me explain how the data is organized: > > > ReferencesToClassLoader.html shows every reference to > [EMAIL PROTECTED] url=null > ,addedOrder=50}, on this case the ClassLoader responsible for deploying > testbyvalue.ear. > > > I have limited this analysis up to 10 levels. A better approach would be dump all objects and their references out to a file. Then write something that runs a full gc() over that information. Once that is done, you need something that lets you "walk the tree". I can imagine a visitor pattern that understands what is permenantly deployed and what is hotdeployed and determines the root cause. Otherwise, a simple GUI would do the trick. -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] RE: FW: JBAS-2972 - OOME / Redeployment leakages
> I don't even know what ThreadPool changes you are referring to, are they real or proposed? I thought we had some changes on the ThreadPools. My bad. I meant that because of references through ThreadWithAttributes (they will be always appear for some reason on JVMTI). Anyway, I need to know if a Thread is cleared (Thread.setContextClassLoader(somethingElse) before the Thread goes to sleep. If the Thread keeps the ClassLoader reference until next request, that will make my life harder creating a TestCase structure for testing redeployment/classLeakages. http://jira.jboss.org/jira/browse/JBAS-2909 -Original Message- From: Adrian Brock Sent: Thursday, March 23, 2006 3:41 AM To: Clebert Suconic Cc: Scott M Stark; jboss-development@lists.sourceforge.net Subject: Re: FW: JBAS-2972 - OOME / Redeployment leakages On Wed, 2006-03-22 at 21:03 -0600, Clebert Suconic wrote: > Sorry... pressed Ctrl-enter instead of Enter. (Damn Outlook) > > Do you think that our changes on ThreadPool would eventually leak > redeployments on the testsuite? (for instance, by keeping > Thread.currentThread().contextClassLoader() still alive) ? Clebert, can you please include some hard facts. I am not going to comment on speculation. If you want me to "get the facts" - (and I don't mean MSFacts ;-) then tell me. I don't even know what ThreadPool changes you are referring to, are they real or proposed? -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] jboss-3.2-testsuite Build Completed With Testsuite Errors
View results here -> http://cruisecontrol.jboss.com/cc/buildresults/jboss-3.2-testsuite?log=log20060323081050 TESTS FAILEDAnt Error Message: /services/cruisecontrol/work/scripts/build-jboss-common.xml:235: The following error occurred while executing this line: /services/cruisecontrol/work/scripts/build-common-targets.xml:26: Build Successful - Tests completed with errors or failures.Date of build: 03/23/2006 08:10:50Time to build: 48 minutes 29 seconds Unit Tests: (1851) Total Errors and Failures: (1)testUnackedQueueorg.jboss.test.jbossmq.test.UnackedUnitTestCase Modifications since last build: (first 50 of 0)
[JBoss-dev] Re: FW: JBAS-2972 - OOME / Redeployment leakages
On Wed, 2006-03-22 at 21:03 -0600, Clebert Suconic wrote: > Sorry... pressed Ctrl-enter instead of Enter. (Damn Outlook) > > Do you think that our changes on ThreadPool would eventually leak > redeployments on the testsuite? (for instance, by keeping > Thread.currentThread().contextClassLoader() still alive) ? Clebert, can you please include some hard facts. I am not going to comment on speculation. If you want me to "get the facts" - (and I don't mean MSFacts ;-) then tell me. I don't even know what ThreadPool changes you are referring to, are they real or proposed? -- Adrian Brock Chief Scientist JBoss Inc. --- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 ___ JBoss-Development mailing list JBoss-Development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jboss-development