You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by "Alexander Martens (JIRA)" <ji...@apache.org> on 2008/10/05 12:24:52 UTC
[jira] Created: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructorSegmentation
activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructorSegmentation
--------------------------------------------------------------------------------------------------
Key: AMQCPP-200
URL: https://issues.apache.org/activemq/browse/AMQCPP-200
Project: ActiveMQ C++ Client
Issue Type: Bug
Affects Versions: 2.2.1
Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
Reporter: Alexander Martens
Assignee: Timothy Bish
Attachments: testConstructor.log.gz
This unit test seems to fail.
I send you attached a gdb postmortem log (bt, bt full, info threads).
Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.9.tar.gz
Good whatever there!
Didn't make it. Stuck on loop #57 (I think we were here already).
I send you attached the gdb's log (thanks to Teemu Torma I've spent some typing:)
info threads
bt
bt full
thread apply all bt
thread apply all bt full
I do also include my local version of OpenwireCmsTemplateTest.h and the start and stop scripts (I know about the needless to export the PYTHONPATH,... shouldn't be a problem, however).
I'm rerunning this time. (I feel curious about the iteration #).
Before you ask, no, no error messages on the server's log.
Read you later!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46197#action_46197 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Odd.
Thanks again Mr Bish.
Will try to update gcc, though I'm running out of ideas.
Don't hesitate to ask for logs / tests.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46886#action_46886 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Oops, well its a start anyway, thanks for keeping up with this. we will solve it eventually. For some reason you seeem to be able to replicate the timing that is causing this deadlock and so far I can't.
At least it looks like my changes have reduced the number of threads to examine which should help.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46418#action_46418 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Alexander, could you please try the attached patch. Using this change in the patch I am able to run all unit tests without any failures on CentOS 5 with APR 1.3.3.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46648#action_46648 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I don't think that this is actually a problem with the client code, but instead a bug in the test. If you look at the source for that testBasics test in the CMSTemplate integration test the code uses a Thread::sleep to wait for another thread to be ready before it starts sending, this is not the best of ideas. I'd bet that if you run the tests enough times that sleep would fail to wait long enough and the receiver would not get all the messages its expecting. I'll have to look at it closer, I didn't write the CMSTemplate stuff so I need to spend some time getting familiar with all the internals but at first glance the sleep makes me very suspicious.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: openwire_transaction.tar.gz
A couple more like deadlock2.
Doesn't want to crash this afternoon - will try at home again.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, openwire_transaction.tar.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: daemons.1.tar.gz
Wow.
Your latest comment has open hell's doors and a couple of daemons came out,... (listening to Haggard on my mp3)
I've commented everything but the CmsTemplate test and added the persistent=false at the "broker" element in my activemq.xml.
Got one segmentation fault. I changed long_run to include a "ulimit -c unlimited" and got one segmentation fault (again) and a few iterations after, a lock.
I send you attached the output, my new activemq.xml, and postmortems for the dumped core and the running deadlock.
Best of all is that I don't need too much time to reproduce it. Is fast!
Hope this helps.
BTW: Segfault becasue of
Program terminated with signal 11, Segmentation fault.
#0 0x001ce5b4 in run_cleanups (cref=0x9fafcf0) at memory/unix/apr_pools.c:2306
2306 (*c->plain_cleanup_fn)((void *)c->data);
* 1 process 3620 0x001ce5b4 in run_cleanups (cref=0x9fafcf0) at memory/unix/apr_pools.c:2306
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46243#action_46243 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
32 bits RHEL5.2 (upgraded from a 5.1 DVD)
Tonight's test on SuSE (64) is because it's a system we already had installed and testing there is just compiling,...
(I absolutely understand what being constrained means, no need to explain that.)
I wish I could let you a quicker setup hint, sorry.
I'm also considering a fresh CentOS 5 install on a brand-new machine (yes, I already did the RHEL5.2 fresh install).
Read you tomorrow!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46441#action_46441 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Sorry, our two comments have entered at a time (Murphy's law).
On my way to test without shared memory now.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46569#action_46569 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
If you are up for some testing I think I have a working fix (fingers crossed). The code is in a branch of activemq-cpp located here:
http://svn.apache.org/repos/asf/activemq/activemq-cpp/branches/pools-revamp
I've run the unit tests, example and integration tests numerous times without any failures. Currently I'm running with everything statically linked but I'd also like to see if the dynamic link also works, it should.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (AMQCPP-200) activemq-cpp-2.2.1 test
fails for decaf::io::ByteArrayInputStreamTest::testConstructor
(segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46539#action_46539 ]
alex.martens edited comment on AMQCPP-200 at 10/16/08 11:17 PM:
---------------------------------------------------------------------
I was fearing something like that. I'm still surprised though, by the fact that this was only detected on CentOS / RHEL.
Count on me if you ever need some more testing!
Thanks again for you effort.
was (Author: alex.martens):
I was fearing something like that. I'm still surprised though, by the fact that this only detected on CentOS / RHEL.
Count on me if you ever need some more testing!
Thanks again for you effort.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46536#action_46536 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I think I may be onto something with this. There's some potential issues with our threads class and the APR pools scheme. Not sure yet how I'm going to fix this one, will require some thought.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.7.log.gz
Sorry for the late answer, it has been a busy day.
Here you're. 20 threads, some reading, some waiting for a condion,...
I've dumped the standard backtrace for each one hoping this is enough for you get a light bulb over your head. If not, the process is stopped in a dedicated server for this issue, so that I can attach gdb and inspect any variable in any thread, any time. Just drop me a mail or post a message.
Hope this helps!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47046#action_47046 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
(New segmentation faults reported at the second sub-task)
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46577#action_46577 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Ok, I will start right now (what are otherwise sundays for?)
Will let you know my results asap.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46890#action_46890 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Done.
Your explanation sounds pretty reasonable to me. If so, this long test is the good one (crossing fingers) - I better go home (is quite late here) and have some dinenr.
(I was wondering... if server quotas are fixed - I didn't change the activemq.xml - as well as the number of messages sent by the test, shouldn't we seeing a stop at the same iteration #, given that my scripts to start and stop activemq - that I'm running at the beginning of each long run - are recursively deleting ActiveMQ's data directory? Hmmm.
Never mind. We'll see thta tomorrow.)
I know I've said this many times, but thanks a lot Timothy.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: testString-O0.log.gz
I've downloaded the latest trunk version from the svn server to see if this issue was already solved, but no luck with it.
If the testConstructor unit test in commented out, a new segmentation fault takes place at
decaf::io::DataInputStreamTest::testString
This time, I've enabled debugging in apr and apr-util, in order to get some backtrace info from apr_pool_cleanup_kill.
Again, hope this helps.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46439#action_46439 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Ok, here's an oddity. Trying building with shared libraries disabled. This is done by running configure as shown below:
{noformat}
configure --enable-shared=no
{noformat}
I had shared libraries turned off when I was debugging and when I made the change in AprPool that was when everything started working for me. See if this happens for you as well. If I turn on shared here it starts failing again for me as well.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47011#action_47011 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
While we've reduced the load by turning off persistence for as many tests as we can, some require it. Also there the fact that we create a large number of topics and queues on each run that are not deleted yet since we don't yet have that feature. If you connect to the broker via JMX after about 30 runs of the client integration tests without restarting the broker you will see a huge list of them. Assuming that the JVM heap space hasn't been set to some huge number I'd expect that eventually we would just exhaust the available resources that the broker has access to.
Someday we will be able to clean up the mess on the broker at then end of each test, but for now we can't.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46191#action_46191 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
fc-9? gcc version?
Anybody running RHEL5.2 ?
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46181#action_46181 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Finished recompiling and retesting with apr and apr util 1.2.12.
Excepting the shared memory test in the apr test unit (shown above), no big deal.
It still crashes at decaf::io::ByteArrayInputStreamTest::testConstructor.
Next unit tests to fail are:
decaf::io::DataInputStreamTest::testUTF
decaf::io::DataInputStreamTest::test_readFullyNullArray
decaf::io::DataInputStreamTest::test_readFullyNullStream
decaf::io::DataInputStreamTest::test_readFullyNullStreamNullArray
decaf::lang::ThreadTest::testDelegate
There are probably more out there, but I need some sleep (it's pretty late here).
I'll try the 1.2.8 rpm package tomorrow in order to discard an apr compilation issue and let you know.
Don't hesitate to ask me for a try with any particular activemq-cpp or apr version.
Thanks for your help!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46999#action_46999 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Alexander
Just wanted to confirm, are you stopping and starting the broker between each run of the integration tests? I ran several continuous runs this weekend with my script that starts and stops a broker and haven't been able to cause the deadlock that you are seeing. From the logs I've seen, the deadlocks all look like the client just didn't receive the expected message from the broker and is stopped waiting.
We should probably make them timed receives with long timeouts so that the tests will eventually fail. I will look into that later on today. I did fix a few memory leaks, some in the lib code, most in the unit tests over the weekend, I am continuing to try and get a clean valgrind run of the unit tests just to make sure we aren't losing any resources anywhere.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46922#action_46922 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
:D
Don't forget this started as a CentOS / RHEL specific issue. I'm still wondering if there's something wrong at libpthread level that prevents other distributions / operating systems from crashing. Must be something really chumsy, though.
K, I'll try that next and post you something asap
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46864#action_46864 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hi!
Using a script - nothing special, just a "for ((i=0; i<500; ++i)) do ... done".
Ok, I'll try to reproduce it again and switch over the other threads to find out who is deadlocking with ~ActiveMQSessionExecutor.
Thanks!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46241#action_46241 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Were you using a 64bit version of RHEL or a 32bit version? I don't think much testing has been done with the CPP library on 64bit and am unsure if it would work out of the box on a 64bit OS. Also not sure of how APR handles 64bit either.
I'm a bit time constrained right now so I have not had much time to look into this.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47012#action_47012 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Well explained.
I'll test it again (I'm still concerned about the latest segmentation faults).
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46663#action_46663 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
You're right again. I was trying not having to step into the code,...
Ok, I commented all stomp tests out and increased every sleep to 1500 (which is not a solution, I know; but I don't think a thread that didn't start in more than a second will ever do, actually). Unfortunately, there are many sleeps there.
I'm starting another 500 run test to see what happens. This should take about a day or so.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46613#action_46613 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
The deadlock logs look like its stuck in a send. Currently the most common reason for a send to stall is that we are being producer flow controlled. Are you connecting to a broker that is in use by anyone else? Is this broker local to the machine you are on or in your control? If so can you change the configuration to disable flow control and run some more tests to see if these stalls repeat?
I've had occasional failures on those stomp tests but they are so infrequent that I've never been able to determine if its the stomp code or the tests that are at fault. I'd ignore stomp issues for the moment unless you are planning on using the stomp protocol which I don't recommend as the openwire protocol performs better and offers a more complete set of features.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46331#action_46331 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hi Timothy!
I've done a fresh install from a CentOS 5.2 DVD (took a while until downloaded from torrent) and I can confirm you the segmentation fault at the same unit test that I saw on Red Hat.
You should be able to reproduce it.
PS: I can post you a ks file for automating the installation and a shell script to automate the builds if it helps.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46237#action_46237 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Carefully following the same steps on a FC8, works fine (as you already posted), even with apr-1.3.
Looks like a RHEL5.2 specific issue. If so, CentOS 5 should have the same pb.
Any CentOS 5 out there?
I'm going to try in a SuSE 10.2 (EL) 64 bits, (which has a similar tool chain version) and I'll post back by tomorrow.
BTW, I got the same testshm failure on apr-1.2.12 under SuSE.
Does this help?
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46915#action_46915 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
On my way
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: storage_logs.tar.gz
Hi!
It is said that devil is hidden in details:
Following the output log, iteration #57 started on oct 29 21:39:34 GMT 2008.
On the server side (no, there are no errors nor warnings, but there is one suspicious INFO line, more than one hour later than its previous trace, and only five minutes before #57):
2008-10-29 21:39:29,738 [127.0.0.1:32874] INFO KahaStore - Kaha Store using data directory /usr/local/apache-activemq-5.1.0/data/localhost/tmp_storage
du -h -s /usr/local/apache-activemq-5.1.0/data/localhost/tmp_storage says 50M (a default maximum value?)
However, if the problem is on the server side (because a resource quota leads it to discard messages, which could be the reason as you pointed out that pushes the client to an endless wait), a client restart (killing the integration test only) should stop on the early loops, shouldn't it?
I'm going to test that...
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock2.log.gz
Another deadlock ? Stuck at this point for ca. 10 minutes.
This one in an apr socket function, rather tan a join.
I've noticed that slowing the execution down (e.g. by connecting the debugger) seems to favor a successful result.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46175#action_46175 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
I get the same error without optimizing (-O0).
Back-trace info is now a bit clearer (new log appended).
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46581#action_46581 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Got one unit test failure under stomp (integration tests - integration::connector::stomp::TransactionTest::test).
I'm trying to find out if this a new issue, an old one or a side effect of using unmanaged apr pools.
Looks great, anyway.
Ok, I will point my svn back to the trunk.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46192#action_46192 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I've personally tested on the following platforms without any test failures.
Fedora 8 - GCC 4.1.2
Fedora 9 - GCC 4.3.0
Sun Solaris 10 - Workshop 11 and 12 + GCC 3.4.6
Sun Solaris 8 - Workshop 11
Windows XP - VC++ 2005 and 2008
Another committer is using Mac OS/X 10.X
I've also had reports from users running various Ubunto and SUSE distributions and CertOS as well. So far no one has reported unit test failures.
I've tried a few configuration changes on F9 but so far cannot reproduce the errors you are seeing.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46592#action_46592 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Good morning!
I run standalone unit tests for 500 times, incident free.
I'll try under RHEL today, though I expect the same result.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.log.gz
An example of deadlock (integration::connector::stomp::ExpirationTest::testNotExpired.414921ff-d26a-1f62-46de-37db4b889abe)
info threads, bt, bt full (as always)
Let's see if I get a crash now, ...
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (AMQCPP-200) activemq-cpp-2.2.1 test
fails for decaf::io::ByteArrayInputStreamTest::testConstructor
(segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46910#action_46910 ]
tabish121 edited comment on AMQCPP-200 at 10/30/08 3:29 AM:
---------------------------------------------------------------
Another thing to try is to go to the integration tests TestRegistry.cpp and comment out everything but the CMS Template test and then run your long test, this will eliminate the other tests which I know are not all leaving the broker in a completely clean state on exit. I went through and set some more to use non persistent messaging last night. Until I get a chance to implement the delete destinations from the broker on exit by there probably will always be a bit of sludge left from running these tests.
You can turn off persistence at the broker end:
http://activemq.apache.org/how-do-i-disable-persistence.html
was (Author: tabish121):
Another thing to try is to go to the integration tests TestRegistry.cpp and comment out everything but the CMS Template test and then run your long test, this will eliminate the other tests which I know are not all leaving the broker in a completely clean state on exit. I went through and set some more to use non persistent messaging last night. Until I get a chance to implement the delete destinations from the broker on exit by there probably will always be a bit of sludge left from running these tests.
You can turn off persistence at the broker end:
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46889#action_46889 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
K, I'm stopping the run and recompiling.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46190#action_46190 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Currently I don't know of anyone else seeing this behaviour. All of the unit tests pass on the machine we test on currently. Windows, Fedora Linux and Mac OS/X. I have not been able to make mine segfault as yet.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46615#action_46615 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Morning there (whatever)!
Sorry, I was using default 5.1 xml. You're right, disabling producer's flow control seems to have dramatically reduced (if not removed at all) openwire deadlocks. I've launched a 500 shot integration-test this morning, just to double check that.
If we stick to openwire (and forget about stomp), this version looks pretty to me!
Again, a big thanks for your help Timothy.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: testConstructor-O0.log.gz
gdb output for
info threads
bt
bt full
(disabled optimizations)
Hope this helps
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46445#action_46445 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hmm,...
Do I need to disable shared memory on apr and apr-util as well ?
It's now crashing at decaf::internal::util::ByteArrayAdapterTest::testClear, just a couple of tests further
I'm working on disabling shared memory on apr-1.3
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: more.daemons.tar.gz
Sorry for the late answer - still fighting too many battles at a time.
Yes, your latest changes were (excepting a !closed in the manager) on the integration tests (they've been nearly rewritten). Testing results, are therefore different too (shouldn't be, but they seem to be).
The short version says:
Segmentation fault on iteration #16: dsm->looseMarshal( this, dataStructure, looseOut ); (core.14475)
Segmentation fault on iteration #36: std::size_t available = inputStream->available(); (core.19442)
Locked on iteration #65
I do also include the output.log, server's log (which is interesting this time; some errors since second 1 or so) and gdb outputs...
BTW: I commented stomp tests out again, just to narrow searches.
Hope this helps!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Timothy Bish updated AMQCPP-200:
--------------------------------
Attachment: CentOS-AprPool-Fix-Test.patch
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46937#action_46937 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
The changes I made probably won't make much difference for you, just some minor clean-ups. I'm still baffled by the last stack trace you sent so once I manage to make some sense of what is going on in there maybe I can find another fix.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47007#action_47007 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
No - I'm starting and stopping the server at every long_test (500 loops, from which less than 70 are actually finished), not within each loop (sorry, may be I wasn't clear enough). However, I thought that forcing persistance to false on the server discarded the disk quota limitation.
Do you think the server is loosing messages? AFAIK there's no expiration time set by default,...
Ok, I'm updating and adding server start and+ stop calls at every loop. Let's see if I we can still trap any segmentation fault.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46244#action_46244 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I've started a download of CentOS, once I get it installed in a VMWare instance I will try and get things building and check the unit tests on that OS.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46888#action_46888 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Looking at the logs it now appears like the code isn't shutting down, but instead waiting because the receiver hasn't received the number of messages that its been told it should have. What I think is happening is that the CMS template code is sending message persistently by default, and we can't purge the queues on exit in our tests yet so you are running out of space on the broker.
I'm betting that if we sent the messages with non persistent messaging we wouldn't see this problem. You can add
cmsTemplate.setDeliveryPersistent(false) to the constructors of the sender and receiver class in the file:
/activemq-trunk-cpp/src/test-integration/integration/connector/openwire/OpenwireCmsTemplateTest.h
I will try this later on to see, but I think I just wasn't running my tests through enough iterations with the broker config that I currently have.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46676#action_46676 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Something that I'm doing... huhh... right... hmmmm.. NOW!
As told, I'll let the test running tonight. Will post something tomorrow.
Thanks Timothy!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46669#action_46669 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Oops!
Got the integration::connector::openwire::OpenwireCmsTemplateTest::testBasics deadlock again (during iteration #60), with 1.5 seconds sleeps.
I'm stopping my tests by now - will continue tonight with a dedicated server and 5 seconds sleeps.
I hope this is not due to a handle leak preventing a new thread from starting.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47031#action_47031 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I believe the first error is related to the issue [AMQ-1712] which should be fixed in the 5.2.0 release. The second one is a result of the durable test, I might at some point be able to stop that from happening but as it stands its not a real issue to be worried about.
Glad to hear we finally broke the 57th iteration.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46872#action_46872 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
That's a lot of threads. I've updated the integration tests so that the CPPUnit framework cleans up in a more timely manner so that there's less stragglers lying around. If you use them we might get a better view of the state of things when the error occurs.
I also made one little change in the Session Executer but I doubt that will fix anything.
Are you running optimized code?
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46583#action_46583 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
I've been running the integration tests and I've seen weird things on both, ActiveMQ 5.1 and 5.3 (snapshot).
I don't find a common pattern nor a deterministic way to reproduce a problem, though issues come pretty easily up.
I'm seeing deadlocks (integration::connector::openwire::OpenwireTransactionTest::test), cms exceptions (StompConnectionNegotiator::requestConnection Negotiate timeout: peer did not send a connected command) even java crashes (I can post the logs, if you want to have a look at them).
Furthermore, I even got the tests to pass without problems from time to time.
Concerning the unit-tests (src/test/activemq-test), I've run them about a 100 consecutive times without problems. I'll let them run a few hundreds more. just to be sure.
(It's close to midnight here, time for me to have some sleep.)
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46910#action_46910 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Another thing to try is to go to the integration tests TestRegistry.cpp and comment out everything but the CMS Template test and then run your long test, this will eliminate the other tests which I know are not all leaving the broker in a completely clean state on exit. I went through and set some more to use non persistent messaging last night. Until I get a chance to implement the delete destinations from the broker on exit by there probably will always be a bit of sludge left from running these tests.
You can turn off persistence at the broker end:
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46891#action_46891 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Not sure if its the real culprit or not, although its possible we have more than one problem. In any case the mode should have been set to non-persistent for this test, so I've updated the code in trunk with this change.
At least we are getting some test code cleanup with all this bug hunting :)
If you want to attach you run script that would give me something else ensure I'm running my tests in a manner that is close to what you are doing so I have a better chance of reproducing the problem.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.6.tar.gz
Hi there,
Again on (integration::connector::openwire::OpenwireCmsTemplateTest::testBasics), on iteration #46 this time.
No changes on the activemq.xml file.
(Remember I've commented all stomp tests out of TestRegistry.cpp)
Hope this helps.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Jacob Rideout (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46946#action_46946 ]
Jacob Rideout commented on AMQCPP-200:
--------------------------------------
I don't know if you resolved this, but here is what we've found with the apr_pools problem:
There is a race condition in
decaf/lang/System.cpp:
AprPool& System::getAprPool() {
static AprPool aprPool;
return aprPool;
}
the static AprPool is being created twice, if two threads call this function at the same time.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46579#action_46579 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
The fixes that have been made so far are now in trunk. I have not found any other issues yet, let me know how your testing goes.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (AMQCPP-200) activemq-cpp-2.2.1 test
fails for decaf::io::ByteArrayInputStreamTest::testConstructor
(segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46941#action_46941 ]
alex.martens edited comment on AMQCPP-200 at 10/31/08 12:00 PM:
---------------------------------------------------------------------
Sorry for the late answer - still fighting too many battles at a time.
Yes, your latest changes were (excepting a !closed in the manager) on the integration tests (they've been nearly rewritten). Testing results, are therefore different too (shouldn't be, but they seem to be).
The short version says:
Segmentation fault on iteration #16: dsm->looseMarshal( this, dataStructure, looseOut ); (core.14475)
Segmentation fault on iteration #36: std::size_t available = inputStream->available(); (core.19442)
Locked on iteration #65
I do also include the output.log, server's log (which is interesting this time; some errors since minute 1 or so) and gdb outputs...
BTW: I commented stomp tests out again, just to narrow searches.
Hope this helps!
was (Author: alex.martens):
Sorry for the late answer - still fighting too many battles at a time.
Yes, your latest changes were (excepting a !closed in the manager) on the integration tests (they've been nearly rewritten). Testing results, are therefore different too (shouldn't be, but they seem to be).
The short version says:
Segmentation fault on iteration #16: dsm->looseMarshal( this, dataStructure, looseOut ); (core.14475)
Segmentation fault on iteration #36: std::size_t available = inputStream->available(); (core.19442)
Locked on iteration #65
I do also include the output.log, server's log (which is interesting this time; some errors since second 1 or so) and gdb outputs...
BTW: I commented stomp tests out again, just to narrow searches.
Hope this helps!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46342#action_46342 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
That's a great new!!
Thanks again Timothy.
I'll be there this weekend, too.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (AMQCPP-200) activemq-cpp-2.2.1 test
fails for decaf::io::ByteArrayInputStreamTest::testConstructor
(segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46187#action_46187 ]
alex.martens edited comment on AMQCPP-200 at 10/6/08 1:29 AM:
-------------------------------------------------------------------
Back there.
I've done a deep clean up this morning and reinstalled apr and apr-util 1.2.8 directly from Apache's rpms (rpm for apr, src.rpm for apr-util since the rpm holds a dependence for libpq 3, which I don't have - using libpq 4 instead; anyway, I don't think this is related to our issue).
The unit test still fails for decaf::io::ByteArrayInputStreamTest::testConstructor.
I've also repeated the whole process on a fresh installed machine (RHEL 5.2 as well), and I'm getting the same result.
Can somebody please confirm the segmentation fault?
Was anyone able to reproduce the crash?
Let me know if you need more information.
Thanks!
was (Author: alex.martens):
Back there.
I've done a deep clean up this morning and reinstalled apr and apr-util 1.2.8 directly from Apache's rpms (rpm for apr, src.rpm for apr-util since the rpm holds a dependence for libpq 3, which I don't have - using libpq 4 instead; anyway, I don't thing this is related to our issue).
The unit test still fails for decaf::io::ByteArrayInputStreamTest::testConstructor.
I've also repeated the whole process on a fresh installed machine (RHEL 5.2 as well), and I'm getting the same result.
Can somebody please confirm the segmentation fault?
Was anyone able to reproduce the crash?
Let me know if you need more information.
Thanks!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46407#action_46407 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
I have to try that.
Funny!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.openwire.join.tar.gz
Bad news. Got a deadlock in an apr_join under openwire after 22 runs.
(integration::connector::openwire::OpenwireCmsTemplateTest::testBasics)
I send you attached gdb's postmortem, my current activemq.xml configuration file and the standard output (which holds additional timing information).
Yes, the server is in my local machine, I have the root password and full control (I'm the only user at this time). No other applications were running with that message broker when testing. I have access to additional machines and linux distributions, if you want me to run additional tests.
You're getting closer, though,...
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46179#action_46179 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Sure, right now.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47020#action_47020 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
K
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: setup.tar.gz
0) Unless you like french keyboards and language, change ks_sup1.cfg under "# Parametres systeme" to your favorite locale.
1) Copy the ks file to an usb pen drive's root and let VMWare have access to your DVD drive and usb ports.
2) On reboot type linux ks=hd:sda1:ks_config.cfg
Perform the installation (should be fairly fast) and reboot.
(Ignore the missing RH package warning, as it is not needed by CentOS)
3) Log in as root, and manually install e2fsprogs-devel (uuid dependency).
mount /dev/cdrom /media
rpm -i /media/CentOS/e2fsprogs-devel-1.39-15.el5.i386.rpm
(Ignore signatures warnings)
4) Create a build directory and a source "packages" one inside it.
mkdir -p /root/builds/packages
5) Copy "clean.sh" to the "builds" directory and the following tarballs into "builds/packages":
activemq-cpp-2.2.1-src.tar.gz
apr-1.3.3.tar.bz2
apr-util-1.3.4.tar.bz2
cppunit-1.12.1.tar.gz
(You don't have to download these from the Internet, do you?)
6) ./clean.sh (time for coffee ** not decaf ** and piece of cake)
7) tail -n 11 activemq-cpp-2.2.1/check.log
---
You should have a configuration.log, build.log and install.log in every directory if you want to review them.
If you prefer to test again apr 1.2, changing "clean.sh" should be really easy,....
The "clean.sh" script takes 18:16 minutes to complete in this machine (might be more under VMWare)
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Timothy Bish resolved AMQCPP-200.
---------------------------------
Resolution: Fixed
Resolved in Trunk.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46892#action_46892 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Sure - I'll post them tomorrow (I'm at home now).
Don't expect any particular stuff:
i) start has a ulimit increasing the number of handlers, exports JAVA_HOME, changes to /usr/local/apache-activemq-xxx, nohup bin/activemq >& /dev/null and launches tail data/activemq.log in background
ii) stop does something similar with activemq-admin stop --all, killall tail and finally rm -rf data/*
Glad to be helpful!
Night.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46422#action_46422 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hi!
Looks like it is still crashing.
I'm doing a clean build in both, RHEL5.2 and CentOS to double check it.
Will try to post the result this morning (CEST).
Any particular apr version do you want me to try with?
Thanks for your interest!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Teemu Torma (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46248#action_46248 ]
Teemu Torma commented on AMQCPP-200:
------------------------------------
Earlier today I had to track down a problem on Solaris where any program got a segfault on exit. The crash happened always in apr_pool_destroy. This may be related to this problem, I am not sure.
The crash is caused by the sequence of events on exit in my case. decaf::internal::AprPool::getAprPool() has a static AprRuntime instance to initialize/terminate apr. The apr terminate will destroy the global apr pool and all it's children, including any pool created by decaf::internal::AprPool. However, the activemq-cpp instance of AprPool is destroyed after AprRuntime and at that point the contained apr_pool is already destroyed and invalid, leading to segfault in apr_pool_destroy.
I have not seen this on Linux, and I guess that the destruction sequence involving method static instances is a very compiler and/or platform dependent.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46578#action_46578 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Timothy,
This branch is looking fine to me. I've been testing it for a while and it seems to pass every unit-test once and again. I'll let it working all night long and I'll perform further testing at the office tomorrow (under RHEL, since I've got only access to Vista and CentOS today).
Congratulations!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47016#action_47016 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Those are troubling yes. I actually did see one or two this weekend but of course wasn't in the debugger when it happened. I've created two subtasks of this issue to separate out the work for seg fault vs deadlocks since we seem to have solved the initial unit test failure problem. If you manage to capture a backtrace on the segfaults or get any new logs for deadlocks can you attach them to those subtasks so we can track the progress on each, thanks.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46862#action_46862 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
This appears to be a deadlock issue on close down. I can't reproduce it here, I've let this test run for hours. How are you running your tests, using a script to continuously start the integration tests, or by modifying the test's main method?
I can't see from the log file what the other threads are doing, it would be nice to figure out who is holding the lock that is causing the deadlock. I'll keep trying.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46873#action_46873 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Morning there!
Yes, I was building with no additional parameters.
Ok, I'll get the latest version, switch optimizations off and recompile the whole thing.
Will post something back asap.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46604#action_46604 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hi!
Sorry I couldn't write before - I hate Mondays.
As expected, RHEL behaves the same way for both, unit tests and integration tests.
I'm running again 5.1 here and the latest I saw was a deadlock,...
Ok, let's see if I get a backtrace...
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46340#action_46340 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Thanks for the contributions. I was able to get the system installed and get all dependencies from RPMForge. I ran a build of the client this morning and am able to reproduce the errors. I've installed my normal Eclipse IDE and other dev tools so this weekend I can begin to try and track down the problem.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46187#action_46187 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Back there.
I've done a deep clean up this morning and reinstalled apr and apr-util 1.2.8 directly from Apache's rpms (rpm for apr, src.rpm for apr-util since the rpm holds a dependence for libpq 3, which I don't have - using libpq 4 instead; anyway, I don't thing this is related to our issue).
The unit test still fails for decaf::io::ByteArrayInputStreamTest::testConstructor.
I've also repeated the whole process on a fresh installed machine (RHEL 5.2 as well), and I'm getting the same result.
Can somebody please confirm the segmentation fault?
Was anyone able to reproduce the crash?
Let me know if you need more information.
Thanks!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46920#action_46920 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
You're seriously making me regret using APR :) The pools thing is extremely annoying.
I don't have time today to look into this to much. I'd suggest trying opposite test now, comment out the CMSTemplate test and run the integration tests to see if its only the CMSTemplate stuff that is mis-behaving. We need to nail down who the real culprit is here.
Looking at the logs I'm not sure what is happening, its possible CMSTemplate is destroying things in a bad way, not sure yet until I dig into the code.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46936#action_46936 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Updated, compiling tests, ...
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46403#action_46403 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Strangely the tests that fail first are the one's that are testing for exceptions to be thrown.
{noformat}
ByteArrayInputStream nullStream;
CPPUNIT_ASSERT_THROW_MESSAGE(
"Should have thrown an IO Exception",
nullStream.read(),
IOException );
{noformat}
Comment that out and the test passes. If you take it farther and go to Exception.cpp and comment out the code in the buildMessage method then this type of test starts passing.
{noformat}
// Allocate buffer with a guess of it's size
AprPool pool;
// Allocate a buffer of the specified size.
char* buffer = apr_pvsprintf( pool.getAprPool(), format, vargs );
// Guessed size was enough. Assign the string.
message.assign( buffer, strlen( buffer ) );
{noformat}
Of course then other tests start failing in even more bizarre places, not not always the same one's. Very strange.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46387#action_46387 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Still scratching my head on this one. For some reason the apr_thread_mutex_destroy call fails because the apr pool that it came from has gotten into a bad state, but how that is happening is still a mystery to me.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46597#action_46597 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Sounds promising. I'm not sure yet about the integration tests, I haven't seen a failure here, but I haven't had time to stress those tests however all my unit test runs have gone without incident except for some random failures in the socket tests which is due to quirks in the VMWare network layer not using addresses quickly enough.
If you can isolate an integration test that is failing and get a back trace on the fail that could be useful.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46178#action_46178 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
Is there anyway you can try this using the 1.2.XX versions of APR. I've not seen this sort of error on the 1.2.XX versions of APR.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: deadlock.8.log.gz
Here you're!
Eight threads, -O0, again in basic on iteration 57.
info threads
bt
bt full
"for each thread from 1 to 8"
thread #i
bt
bt full
(I've manually added some carry returns to make it more readable)
I'll keep the process running - don't hesitate to ask for a variable / watch.
Best,
Alex
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Summary: activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault) (was: activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructorSegmentation)
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46675#action_46675 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I've removed the reliance on the sleep method from this test so you shouldn't have to muck with the sleep time value any more if you grab the latest code from trunk.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46442#action_46442 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I'm aware of the per 1.3 issue, the apr_pool_create_core method is a new. If this works I will fix it so that it calls the deprecated version when used with the older libraries.
In my tests the when using shared libraries with the change I made I am seeing failures in all kinds of different tests randomly, I even had them pass once.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46608#action_46608 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Got this one after some (4) successfull runs.
integration::connector::stomp::SimpleTest::testMultipleConnections. : assertionF
integration::connector::stomp::SimpleTest::testMultipleSessions. : OK
integration::connector::stomp::SimpleTest::testReceiveAlreadyInQueue. : OK
integration::connector::stomp::TransactionTest::test. : OK
integration::connector::stomp::StompStressTests::testRapidCreate. : OK
!!!FAILURES!!!
Test Results:
Run: 41 Failures: 1 Errors: 0
1) test: integration::connector::stomp::SimpleTest::testMultipleConnections (F) line: 371 integration/connector/stomp/SimpleTest.cpp
assertion failed
- Expression: message != NULL
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46446#action_46446 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I did not disable the shared APR libraries so we are still linking against them if they are there. I did just see a test failure finally. So yes the next thing would be to disable the shared apr libs so we link statically with it.
Not sure at this point what the issue would be with static vs shared libraries.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46885#action_46885 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Dammit!
I've just realized that I forgot to add -g to my CFLAGS, CXXFLAGS and therefore this log has call stack and no optimizations, but no local variable information for apr / apr-util and activemq-cpp. Sorry, my fault!
I'm rebuilding and rerunning.
Will post something back tomorrow morning.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Timothy Bish updated AMQCPP-200:
--------------------------------
Fix Version/s: 2.2.2
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: quota.2.tar.gz
Well, hmmm, it's not our luckiest day ever, but we're getting closer.
This one hung up on iteration #63, after nearly 4 hours running, on
integration::connector::openwire::OpenwireTempDestinationTest::test
The tarball contains gdb's log and the output (timing purposes).
Nothing on the server log, not even a single INFO line - beside teh start bunch.
Time for me to go home. I'll keep the process stopped, just in case you need extra information.
:?
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46539#action_46539 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
I was fearing something like that. I'm still surprised though, by the fact that this only detected on CentOS / RHEL.
Count on me if you ever need some more testing!
Thanks again for you effort.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Attachment: seg_fault.log.gz
A segmentation fault when testing rapid stomp creation
integration::connector::stomp::StompStressTests::testRapidCreate
Hope this one helps!
BTW, this one fails from time to time:
1) test: integration::connector::stomp::SimpleTest::testMultipleConnections (F) line: 371 integration/connector/stomp/SimpleTest.cpp
assertion failed
- Expression: message != NULL
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, deadlock.log.gz, deadlock2.log.gz, openwire_transaction.tar.gz, seg_fault.log.gz, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46180#action_46180 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Looks like there's something weird with apr-1.2.12, when compiled with -O0
testatomic : SUCCESS
testdir : SUCCESS
testdso : SUCCESS
testdup : SUCCESS
testenv : SUCCESS
testfile : SUCCESS
testfilecopy : SUCCESS
testfileinfo : SUCCESS
testflock : SUCCESS
testfmt : SUCCESS
testfnmatch : SUCCESS
testargs : SUCCESS
testhash : SUCCESS
testipsub : SUCCESS
testlock : SUCCESS
testlfs : SUCCESS
testmmap : SUCCESS
testnames : SUCCESS
testoc : SUCCESS
testpath : SUCCESS
testpipe : SUCCESS
testpoll : SUCCESS
testpools : SUCCESS
testproc : SUCCESS
testprocmutex : SUCCESS
testrand : SUCCESS
testrand2 : SUCCESS
testsleep : SUCCESS
testshm : FAILED 1 of 6
testsock : SUCCESS
testsockets : SUCCESS
testsockopt : SUCCESS
teststr : SUCCESS
teststrnatcmp : SUCCESS
testtable : SUCCESS
testtemp : SUCCESS
testthread : SUCCESS
testtime : SUCCESS
testud : SUCCESS
testuser : SUCCESS
testvsn : SUCCESS
Failed Tests Total Fail Failed %
===================================================
testshm 6 1 16.67%
This didn't happen with apr-1.3
Do you want me to continue?
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46297#action_46297 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Kiitos Teemu!
I'm debugging in this direction as well (because of another issue), though I'm not sure this is related to a probably broken build in RHEL5.2
We'll see.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Timothy Bish (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46332#action_46332 ]
Timothy Bish commented on AMQCPP-200:
-------------------------------------
I beat you to it :)
I am working on getting my development environment set up so I can work on debugging the problem.
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47028#action_47028 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Heading home.
I've taken a quick look at the test and no locks nor segmentation faults were detected at this point (loop #58), which is a good new.
Let's see if it continues this way until tomorrow.
The only remarkable thing is a set of errors in the server log (two types: "Subscription does not exist" and "Channel inactive for too long"). I think the second one does only appear when shutting teh server down. (I can append the log if you want, though I'd say you're seeing something similar as our configuration files are almost identical.)
Read you tomorrow!
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Fix For: 2.2.2
>
> Attachments: CentOS-AprPool-Fix-Test.patch, daemons.1.tar.gz, deadlock.6.tar.gz, deadlock.7.log.gz, deadlock.8.log.gz, deadlock.9.tar.gz, deadlock.log.gz, deadlock.openwire.join.tar.gz, deadlock2.log.gz, more.daemons.tar.gz, openwire_transaction.tar.gz, quota.2.tar.gz, seg_fault.log.gz, setup.tar.gz, storage_logs.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46440#action_46440 ]
Alexander Martens commented on AMQCPP-200:
------------------------------------------
Hi there.
My test results are (after applying the patch and performing clean builds):
RHEL5.2, apr 1.3.3 + apr-util 1.3.4:
Unit tests fail randomly with segmentation faults.
RHEL5.2, apr 1.2.8 (rpm) and apr-util 1.2.8 (src.rpm)
Doesn't compile (apr_pool_create_core is not there. "Grepping" this symbol in /usr/lib/apr-1 shows no results)
CentOS5.2, apr 1.3.3 + apr-util 1.3.4:
Like RHEL, random segmentation faults.
In three different tries:
decaf::internal::nio::ShortArrayBufferTest::testGetWithIndex/bin/sh: line 4: 18092 Erreur de segmentation ${dir}$tst
decaf::internal::util::ByteArrayAdapterTest::testReszie/bin/sh: line 4: 18278 Erreur de segmentation ${dir}$tst
decaf::internal::util::ByteArrayAdapterTest::testWriteExceptions/bin/sh: line 4: 18463 Erreur de segmentation ${dir}$tst
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructor (segmentation fault)
> -----------------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: CentOS-AprPool-Fix-Test.patch, setup.tar.gz, testConstructor-O0.log.gz, testConstructor.log.gz, testString-O0.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (AMQCPP-200) activemq-cpp-2.2.1 test fails for
decaf::io::ByteArrayInputStreamTest::testConstructorSegmentation
Posted by "Alexander Martens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/activemq/browse/AMQCPP-200?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Martens updated AMQCPP-200:
-------------------------------------
Component/s: Decaf
> activemq-cpp-2.2.1 test fails for decaf::io::ByteArrayInputStreamTest::testConstructorSegmentation
> --------------------------------------------------------------------------------------------------
>
> Key: AMQCPP-200
> URL: https://issues.apache.org/activemq/browse/AMQCPP-200
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Components: Decaf
> Affects Versions: 2.2.1
> Environment: RHEL 5.2 (2.6.18-92.el5), apr-1.3.3, apr-util-1.3.4, cppunit-1.12.1, gcc version 4.1.2 20071124 (Red Hat 4.1.2-42), Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
> Reporter: Alexander Martens
> Assignee: Timothy Bish
> Attachments: testConstructor.log.gz
>
>
> This unit test seems to fail.
> I send you attached a gdb postmortem log (bt, bt full, info threads).
> Thanks!
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.