[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #8 from Mark Thomas --- Thanks for the confirmations and updating the bug. I was just waiting to complete a full unit test run for 9.0.x, 8.5.x and 7.0.x before updating. They have just finished. All looks good. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 Konstantin Kolinko changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #7 from Konstantin Kolinko --- Mark wrote in "[VOTE] Release Apache Tomcat 7.0.98" thread > Found it. It is a bug in the unit test. The client is closing the > connection too soon. I'm not sure why we are seeing this on a subset of > operating systems and a subset of versions but I'll get it fixed generally. (In reply to Michael Osipov from comment #6) > 609d10230ad1cb67c9745764ccee98450df89880 fixes the issue for me. I confirm as well. That commit fixes the issue for me as well, on Windows 10. I updated the changelog file. I am closing this issue as FIXED. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #6 from Michael Osipov --- 609d10230ad1cb67c9745764ccee98450df89880 fixes the issue for me. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #5 from Michael Osipov --- I just went back in history and even the first compilable version in 7.0.x this test does block forever. Your change does not block forever anymore, but makes the test fail after 10 seconds. So I will always see failed tests. What now? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #4 from Konstantin Kolinko --- Created attachment 36898 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36898=edit Log output - waiting is limited to 10 seconds (858f7e7d) 1. In the following commit I limited the wait time on the latch to 10 seconds https://github.com/apache/tomcat/commit/858f7e7d3d367f4654a95aa31f0ebf57185e12cc The result is that - the following tests are all succeeding Testcase: testAsync[0: end [NONE], timing [INLINE]] took 2,698 sec Testcase: testAsync[1: end [NONE], timing [THREAD_BEFORE_EXIT]] took 1,117 sec Testcase: testAsync[2: end [NONE], timing [THREAD_AFTER_EXIT]] took 2,117 sec Testcase: testAsync[3: end [COMPLETE], timing [INLINE]] took 3,019 sec Testcase: testAsync[4: end [COMPLETE], timing [THREAD_BEFORE_EXIT]] took 3,016 sec Testcase: testAsync[5: end [COMPLETE], timing [THREAD_AFTER_EXIT]] took 4,038 sec Testcase: testAsync[6: end [DISPATCH], timing [INLINE]] took 3,018 sec Testcase: testAsync[7: end [DISPATCH], timing [THREAD_BEFORE_EXIT]] took 3,045 sec Testcase: testAsync[8: end [DISPATCH], timing [THREAD_AFTER_EXIT]] took 4,024 sec - the following tests are all failing Testcase: testAsync[9: end [ERROR_NONE], timing [INLINE]] took 10,115 sec Testcase: testAsync[10: end [ERROR_NONE], timing [THREAD_AFTER_EXIT]] took 10,1 sec Testcase: testAsync[11: end [ERROR_COMPLETE], timing [INLINE]] took 10,099 sec Testcase: testAsync[12: end [ERROR_COMPLETE], timing [THREAD_AFTER_EXIT]] took 10,102 sec Testcase: testAsync[13: end [ERROR_DISPATCH], timing [INLINE]] took 10,099 sec Testcase: testAsync[14: end [ERROR_DISPATCH], timing [THREAD_AFTER_EXIT]] took 10,101 sec 2. Looking into commit history of TestAsyncContextStateChanges.java file, the following commit on 16 Oct 2019 disabled some tests due to known deadlocks. https://github.com/apache/tomcat/commit/91aa7e888c64e6aa4c5bf1625024774dbe8f843e -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 Konstantin Kolinko changed: What|Removed |Added Attachment #36893|Thread dump of locked JVM |Thread dump of locked JVM - description||FreeBSD -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #3 from Konstantin Kolinko --- Created attachment 36897 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36897=edit Captured stdout/stderr of Apache Ant - Windows, Java 13.0.1 -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #2 from Konstantin Kolinko --- Created attachment 36896 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36896=edit Thread dump of locked JVM - Windows, Java 13.0.1 -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 63950] org.apache.catalina.core.TestAsyncContextStateChanges with APR blocks forever on FreeBSD
https://bz.apache.org/bugzilla/show_bug.cgi?id=63950 --- Comment #1 from Konstantin Kolinko --- I am testing release candidate of Tomcat 7.0.98 on Windows 10, and it hangs on the same test class and with the same stacktrace, but - Michael's version of Tomcat Native is different, - a different test case in the test class is hanging. 1. In my case I use Tomcat Native 1.2.23 > [junit] 22.11.2019 23:41:04 org.apache.catalina.core.AprLifecycleListener > lifecycleEvent > [junit] INFO: Loaded APR based Apache Tomcat Native library [1.2.23] using > APR version [1.7.0]. Michael used 1.2.21. 2. I tested - 32-bit native with 32-bit Oracle Java 6u45, Oracle Java 7u80 - 64-bit native with 64-bit Java 8u222 from AdoptOpenJDK, Java 13 from OpenJDK. In all cases the hanging test case is #9: > [junit] INFO: Starting test case [testAsync[9: end [ERROR_NONE], timing > [INLINE]]] For Michael the hanging testcase was #10, as if #9 completed successfully: > [junit] INFORMATION: Starting test case [testAsync[10: end [ERROR_NONE], > timing [THREAD_AFTER_EXIT]]] 3. I do not see any significant differences in the stacktraces. It is a parameterized test, thus line numbers are the same. For NIO, BIO connectors the same test class completes successfully in 13 seconds. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
