[jira] [Created] (QPID-5400) Deadlock when using amqp-1-0 client to connect to amqp-0-9-1 broker
Brian Clark created QPID-5400: - Summary: Deadlock when using amqp-1-0 client to connect to amqp-0-9-1 broker Key: QPID-5400 URL: https://issues.apache.org/jira/browse/QPID-5400 Project: Qpid Issue Type: Bug Components: Java Client Affects Versions: 0.24 Reporter: Brian Clark Reported at thread http://qpid.2158936.n2.nabble.com/Deadlock-when-using-amqp-1-0-client-to-connect-to-amqp-0-9-1-broker-td7601480.html -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Updated] (QPID-5400) Deadlock when using amqp-1-0 client to connect to amqp-0-9-1 broker
[ https://issues.apache.org/jira/browse/QPID-5400?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Brian Clark updated QPID-5400: -- Labels: deadlock (was: ) Deadlock when using amqp-1-0 client to connect to amqp-0-9-1 broker --- Key: QPID-5400 URL: https://issues.apache.org/jira/browse/QPID-5400 Project: Qpid Issue Type: Bug Components: Java Client Affects Versions: 0.24 Reporter: Brian Clark Labels: deadlock Reported at thread http://qpid.2158936.n2.nabble.com/Deadlock-when-using-amqp-1-0-client-to-connect-to-amqp-0-9-1-broker-td7601480.html -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Reopened] (QPID-5397) Dispatch - Crash occurs when linked deliveries are concurrently settled
[ https://issues.apache.org/jira/browse/QPID-5397?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ted Ross reopened QPID-5397: The fix was an improvement, but there is still a window of opportunity for a problem. If a delivery is sent unsettled, and is then settled by the sender before being settled by the receiver (due to exceeding the outgoing window in Messenger for example), Dispatch may attempt to re-settle the sender-side delivery when the receiver settles thus causing problems in Proton Engine and possibly double-freeing the structure. A new fix will hold off the unlinking and freeing of linked deliveries until they are both settled. Dispatch - Crash occurs when linked deliveries are concurrently settled --- Key: QPID-5397 URL: https://issues.apache.org/jira/browse/QPID-5397 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross Dispatch router crashed while running a test in which a client/server pair exchanges requests and responses with acknowledgement (not pre-settled). The test does not throttle the message rates. After several tens-of-thousands of transfers, the router crashed while settling a delivery and attempting to unlink it from its peer. Apparently two linked deliveries were concurrently settled on separate threads. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Created] (QPID-5401) Linearstore: ARM compile error: cast ... increases required alignment of target type
Kim van der Riet created QPID-5401: -- Summary: Linearstore: ARM compile error: cast ... increases required alignment of target type Key: QPID-5401 URL: https://issues.apache.org/jira/browse/QPID-5401 Project: Qpid Issue Type: Bug Affects Versions: 0.26 Reporter: Kim van der Riet Assignee: Kim van der Riet Compiling for ARM on 0.26 exposes several instances of the compile error: {color:blue}{noformat} cast from 'char*' to 'uint64_t* {aka long long unsigned int*}' increases required alignment of target type [-Werror=cast-align] {noformat}{color} or similar in the linearstore. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Created] (QPID-5402) Disable default build of linearstore in cmake
Kim van der Riet created QPID-5402: -- Summary: Disable default build of linearstore in cmake Key: QPID-5402 URL: https://issues.apache.org/jira/browse/QPID-5402 Project: Qpid Issue Type: Bug Affects Versions: 0.26 Reporter: Kim van der Riet Priority: Minor Current cmake configuration will build both the legacystore and linearstore (if the requisite dependent components are present). Until linearstore reaches greater maturity, only the legacystore should be built by default. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Assigned] (QPID-5402) Disable default build of linearstore in cmake
[ https://issues.apache.org/jira/browse/QPID-5402?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Kim van der Riet reassigned QPID-5402: -- Assignee: Kim van der Riet Disable default build of linearstore in cmake - Key: QPID-5402 URL: https://issues.apache.org/jira/browse/QPID-5402 Project: Qpid Issue Type: Bug Affects Versions: 0.26 Reporter: Kim van der Riet Assignee: Kim van der Riet Priority: Minor Current cmake configuration will build both the legacystore and linearstore (if the requisite dependent components are present). Until linearstore reaches greater maturity, only the legacystore should be built by default. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
Re: Review Request 15818: QPID-5307: Make brokertest.py framework run on Windows, convert some tests.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/15818/#review29876 --- /trunk/qpid/cpp/src/qpid/sys/windows/Path.cpp https://reviews.apache.org/r/15818/#comment57341 The separator should be '\\' here. - Chug Rolke On Dec. 4, 2013, 10:20 p.m., Alan Conway wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/15818/ --- (Updated Dec. 4, 2013, 10:20 p.m.) Review request for qpid, Andrew Stitcher, Chug Rolke, Justin Ross, Rafael Schloming, and Steve Huston. Bugs: qpid-5307 https://issues.apache.org/jira/browse/qpid-5307 Repository: qpid Description --- QPID-5307: Make brokertest.py framework run on Windows, convert some tests. The brokertest framework is a python unittest framework designed for test that need to run brokers and other processes. This commit makes the framework portable to window and converts the following test shell/ps1 scripts to python: - run_python_tests - python_tests.py - run_acl_tests - acl_tests.py Each test module is a set of python unittests, and also an executable script that runs itself using qpid/python/qpid-python-test. - No environment setting required, test_env.sh and test_env.ps1 are not required. (Note they are still useful for working with a build from the command line.) - Configuration formerly in the test_env scripts is provided in test_config.py - Only requirement to run test modules is that the current directory: - contains a test_config.py script (cmake creates one in build/src/tests) - is writable so the tests can create output directories. Each BrokerTest stores broker logs, data directories, command line, stderr and stdout of all processes run in test-module.tmp/test-name. The framework automaticaly handles shutting down brokers and killing leftover processes at the end of the test, and verifies that all process exit with the expected exit status (expectation of running/exit-ok/crash can be set for each process.) Some broker tests start a broker and then run other python tests against it (e.g. pyton_tests runs the tests under qpid/python, acl_tests runs the pre-existing acl.py tests.) Those sub-tests don't store output on disk so there's no corresponding directory in the test-module.tmp tree. This is not yet complete, things to do include: - Generate test_config.py to run tests against a qpid install. - Tests failing on windows: - broker crashes during python_tests.PythonTest.test_run_python_tests - acl_tests.AclTests.test_noacl_noauth_deny_create_link assertion fails. - ha_tests hang. - missing store setup specific to windows stores. - Broke pre-existing tests that use brokertest: configuration has moved. - store.py - qpidd_qmfv2_tests.py - interlink_tests.py - legacystore/python_tests/client_persistence.py - legacystore/python_tests/resize.py - legacystore/python_tests/store_test.py - Nested test implementation self.run_test_modules is clunky. - confusing reporting of # tests run. - tests selection patterns need to include the parent test and the sub-tests. QPID-5307: Added portable Path, check for isAbsolute() qpid::sys::Path provides some basic operations on path names, with implementations for posix and windows. QPID-5307: Fix bug in windows NullAuthenticator, not setting realm on user IDs. Diffs - /trunk/qpid/cpp/src/CMakeLists.txt 1546628 /trunk/qpid/cpp/src/qpid/acl/AclPlugin.cpp 1546628 /trunk/qpid/cpp/src/qpid/broker/windows/SaslAuthenticator.cpp 1546628 /trunk/qpid/cpp/src/qpid/sys/Path.h PRE-CREATION /trunk/qpid/cpp/src/qpid/sys/posix/Path.cpp PRE-CREATION /trunk/qpid/cpp/src/qpid/sys/windows/Path.cpp PRE-CREATION /trunk/qpid/cpp/src/tests/CMakeLists.txt 1546628 /trunk/qpid/cpp/src/tests/acl.py 1546628 /trunk/qpid/cpp/src/tests/acl_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/brokertest.py 1546628 /trunk/qpid/cpp/src/tests/ha_test.py 1546628 /trunk/qpid/cpp/src/tests/ha_tests.py 1546628 /trunk/qpid/cpp/src/tests/python_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/run_acl_tests 1546628 /trunk/qpid/cpp/src/tests/run_acl_tests.ps1 1546628 /trunk/qpid/cpp/src/tests/test_config.py.in PRE-CREATION /trunk/qpid/python/qpid-python-test 1546628 Diff: https://reviews.apache.org/r/15818/diff/ Testing --- ha_tests.py passing Thanks, Alan Conway
[jira] [Created] (QPID-5403) Dispatch - The router-specific annotations have reserved keys
Ted Ross created QPID-5403: -- Summary: Dispatch - The router-specific annotations have reserved keys Key: QPID-5403 URL: https://issues.apache.org/jira/browse/QPID-5403 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross The keys in the annotations used in Dispatch Router are reserved. They need to be changed so they are prefixed with x-opt-. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
Re: Review Request 15818: QPID-5307: Make brokertest.py framework run on Windows, convert some tests.
On Dec. 6, 2013, 4:02 p.m., Chug Rolke wrote: /trunk/qpid/cpp/src/qpid/sys/windows/Path.cpp, line 61 https://reviews.apache.org/r/15818/diff/2/?file=394149#file394149line61 The separator should be '\\' here. separator[0] is '\\' - I think the check for '/' also catches the valid leading / - so this code is correct, I believe. - Steve --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/15818/#review29876 --- On Dec. 4, 2013, 10:20 p.m., Alan Conway wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/15818/ --- (Updated Dec. 4, 2013, 10:20 p.m.) Review request for qpid, Andrew Stitcher, Chug Rolke, Justin Ross, Rafael Schloming, and Steve Huston. Bugs: qpid-5307 https://issues.apache.org/jira/browse/qpid-5307 Repository: qpid Description --- QPID-5307: Make brokertest.py framework run on Windows, convert some tests. The brokertest framework is a python unittest framework designed for test that need to run brokers and other processes. This commit makes the framework portable to window and converts the following test shell/ps1 scripts to python: - run_python_tests - python_tests.py - run_acl_tests - acl_tests.py Each test module is a set of python unittests, and also an executable script that runs itself using qpid/python/qpid-python-test. - No environment setting required, test_env.sh and test_env.ps1 are not required. (Note they are still useful for working with a build from the command line.) - Configuration formerly in the test_env scripts is provided in test_config.py - Only requirement to run test modules is that the current directory: - contains a test_config.py script (cmake creates one in build/src/tests) - is writable so the tests can create output directories. Each BrokerTest stores broker logs, data directories, command line, stderr and stdout of all processes run in test-module.tmp/test-name. The framework automaticaly handles shutting down brokers and killing leftover processes at the end of the test, and verifies that all process exit with the expected exit status (expectation of running/exit-ok/crash can be set for each process.) Some broker tests start a broker and then run other python tests against it (e.g. pyton_tests runs the tests under qpid/python, acl_tests runs the pre-existing acl.py tests.) Those sub-tests don't store output on disk so there's no corresponding directory in the test-module.tmp tree. This is not yet complete, things to do include: - Generate test_config.py to run tests against a qpid install. - Tests failing on windows: - broker crashes during python_tests.PythonTest.test_run_python_tests - acl_tests.AclTests.test_noacl_noauth_deny_create_link assertion fails. - ha_tests hang. - missing store setup specific to windows stores. - Broke pre-existing tests that use brokertest: configuration has moved. - store.py - qpidd_qmfv2_tests.py - interlink_tests.py - legacystore/python_tests/client_persistence.py - legacystore/python_tests/resize.py - legacystore/python_tests/store_test.py - Nested test implementation self.run_test_modules is clunky. - confusing reporting of # tests run. - tests selection patterns need to include the parent test and the sub-tests. QPID-5307: Added portable Path, check for isAbsolute() qpid::sys::Path provides some basic operations on path names, with implementations for posix and windows. QPID-5307: Fix bug in windows NullAuthenticator, not setting realm on user IDs. Diffs - /trunk/qpid/cpp/src/CMakeLists.txt 1546628 /trunk/qpid/cpp/src/qpid/acl/AclPlugin.cpp 1546628 /trunk/qpid/cpp/src/qpid/broker/windows/SaslAuthenticator.cpp 1546628 /trunk/qpid/cpp/src/qpid/sys/Path.h PRE-CREATION /trunk/qpid/cpp/src/qpid/sys/posix/Path.cpp PRE-CREATION /trunk/qpid/cpp/src/qpid/sys/windows/Path.cpp PRE-CREATION /trunk/qpid/cpp/src/tests/CMakeLists.txt 1546628 /trunk/qpid/cpp/src/tests/acl.py 1546628 /trunk/qpid/cpp/src/tests/acl_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/brokertest.py 1546628 /trunk/qpid/cpp/src/tests/ha_test.py 1546628 /trunk/qpid/cpp/src/tests/ha_tests.py 1546628 /trunk/qpid/cpp/src/tests/python_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/run_acl_tests 1546628 /trunk/qpid/cpp/src/tests/run_acl_tests.ps1 1546628 /trunk/qpid/cpp/src/tests/test_config.py.in PRE-CREATION /trunk/qpid/python/qpid-python-test 1546628 Diff: https://reviews.apache.org/r/15818/diff/ Testing --- ha_tests.py passing Thanks, Alan Conway
[jira] [Commented] (QPID-5403) Dispatch - The router-specific annotations have reserved keys
[ https://issues.apache.org/jira/browse/QPID-5403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841381#comment-13841381 ] ASF subversion and git services commented on QPID-5403: --- Commit 1548580 from [~tedross] in branch 'dispatch/trunk' [ https://svn.apache.org/r1548580 ] QPID-5403 - Renamed the annotation keys to avoid the reserved name space. Dispatch - The router-specific annotations have reserved keys - Key: QPID-5403 URL: https://issues.apache.org/jira/browse/QPID-5403 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross The keys in the annotations used in Dispatch Router are reserved. They need to be changed so they are prefixed with x-opt-. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Resolved] (QPID-5403) Dispatch - The router-specific annotations have reserved keys
[ https://issues.apache.org/jira/browse/QPID-5403?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ted Ross resolved QPID-5403. Resolution: Fixed Dispatch - The router-specific annotations have reserved keys - Key: QPID-5403 URL: https://issues.apache.org/jira/browse/QPID-5403 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross The keys in the annotations used in Dispatch Router are reserved. They need to be changed so they are prefixed with x-opt-. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
Review Request 16084: [Python] Modify python API to consistenly use timeout values expressed as seconds.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/16084/ --- Review request for qpid and Rafael Schloming. Bugs: proton-475 https://issues.apache.org/jira/browse/proton-475 Repository: qpid Description --- See bug link. This patch unfortunately changes the behavior of the API, and could break existing code. The following methods are affected: Transport.idle_timeout - value now in seconds (was msecs) Transport.remote_idle_timeout - ditto Transport.tick(now) - now changed to seconds since Epoch (was msecs) Driver.wait(timeout) - timeout now in seconds (was msecs) This change is consistent with the traditional use of time and timeouts under Python. Diffs - /proton/trunk/proton-c/bindings/python/proton.py 1548123 /proton/trunk/tests/python/proton_tests/engine.py 1548123 Diff: https://reviews.apache.org/r/16084/diff/ Testing --- unit tests Thanks, Kenneth Giusti
[jira] [Commented] (QPID-5398) qpidd --acl-file does not work with a drive-prefixed path on windows.
[ https://issues.apache.org/jira/browse/QPID-5398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841413#comment-13841413 ] ASF subversion and git services commented on QPID-5398: --- Commit 1548596 from [~aconway] in branch 'qpid/trunk' [ https://svn.apache.org/r1548596 ] QPID-5398: Fix cut 'n paste error in sys/posix/Path.cpp Posix Path::IsAbsolute() function was checking for windows-style paths. qpidd --acl-file does not work with a drive-prefixed path on windows. - Key: QPID-5398 URL: https://issues.apache.org/jira/browse/QPID-5398 Project: Qpid Issue Type: Bug Components: C++ Broker Affects Versions: 0.24 Reporter: Alan Conway Assignee: Alan Conway The acl-file option does not correctly recognize windows paths of the form x:\... as absolute paths and attempts to append them to the brokers data-dir To reproduce (on windows): qpidd --acl-file c:\path\to\policy.acl Result: log contains error Unable to open ACL file data-dir/c:\path\to\policy.acl Expected result: should load file correctly. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
Re: Review Request 15818: QPID-5307: Make brokertest.py framework run on Windows, convert some tests.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/15818/ --- (Updated Dec. 6, 2013, 5:05 p.m.) Review request for qpid, Andrew Stitcher, Chug Rolke, Justin Ross, Rafael Schloming, and Steve Huston. Changes --- Rebased on latest trunk -r1548596 Bugs: qpid-5307 https://issues.apache.org/jira/browse/qpid-5307 Repository: qpid Description (updated) --- QPID-5307: Make brokertest.py framework run on Windows, convert some tests. The brokertest framework is a python unittest framework designed for test that need to run brokers and other processes. This commit makes the framework portable to window and converts the following test shell/ps1 scripts to python: - run_python_tests - python_tests.py - run_acl_tests - acl_tests.py Each test module is a set of python unittests, and also an executable script that runs itself using qpid/python/qpid-python-test. - No environment setting required, test_env.sh and test_env.ps1 are not required. (Note they are still useful for working with a build from the command line.) - Configuration formerly in the test_env scripts is provided in test_config.py - Only requirement to run test modules is that the current directory: - contains a test_config.py script (cmake creates one in build/src/tests) - is writable so the tests can create output directories. Each BrokerTest stores broker logs, data directories, command line, stderr and stdout of all processes run in test-module.tmp/test-name. The framework automaticaly handles shutting down brokers and killing leftover processes at the end of the test, and verifies that all process exit with the expected exit status (expectation of running/exit-ok/crash can be set for each process.) Some broker tests start a broker and then run other python tests against it (e.g. pyton_tests runs the tests under qpid/python, acl_tests runs the pre-existing acl.py tests.) Those sub-tests don't store output on disk so there's no corresponding directory in the test-module.tmp tree. This is not yet complete, things to do include: - Generate test_config.py to run tests against a qpid install. - Tests failing on windows: - broker crashes during python_tests.PythonTest.test_run_python_tests - acl_tests.AclTests.test_noacl_noauth_deny_create_link assertion fails. - ha_tests hang. - missing store setup specific to windows stores. - Broke pre-existing tests that use brokertest: configuration has moved. - store.py - qpidd_qmfv2_tests.py - interlink_tests.py - legacystore/python_tests/client_persistence.py - legacystore/python_tests/resize.py - legacystore/python_tests/store_test.py - Nested test implementation self.run_test_modules is clunky. - confusing reporting of # tests run. - tests selection patterns need to include the parent test and the sub-tests. Diffs (updated) - /trunk/qpid/cpp/src/tests/CMakeLists.txt 1548596 /trunk/qpid/cpp/src/tests/acl.py 1548596 /trunk/qpid/cpp/src/tests/acl_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/brokertest.py 1548596 /trunk/qpid/cpp/src/tests/ha_test.py 1548596 /trunk/qpid/cpp/src/tests/ha_tests.py 1548596 /trunk/qpid/cpp/src/tests/python_tests.py PRE-CREATION /trunk/qpid/cpp/src/tests/run_acl_tests 1548596 /trunk/qpid/cpp/src/tests/run_acl_tests.ps1 1548596 /trunk/qpid/cpp/src/tests/test_config.py.in PRE-CREATION /trunk/qpid/python/qpid-python-test 1548596 Diff: https://reviews.apache.org/r/15818/diff/ Testing --- ha_tests.py passing Thanks, Alan Conway
[jira] [Created] (QPID-5404) HA broker message duplication when deleting a queue with an alt-exchange
Alan Conway created QPID-5404: - Summary: HA broker message duplication when deleting a queue with an alt-exchange Key: QPID-5404 URL: https://issues.apache.org/jira/browse/QPID-5404 Project: Qpid Issue Type: Bug Components: C++ Clustering Affects Versions: 0.24 Reporter: Alan Conway Assignee: Alan Conway Description of problem: When deleting a queue with an alternate exchange, both the primary and the backup brokers will route messages to the alt-exchange when the queue is deleted. This will cause the backups to potentially receive duplicate copies of the message if the alt-exchange is bound to a replicated queue (one set from the local route to the alt-exchange and another from the primary due to its route to the alt-exchange). Suggest that the backup brokers should not route messages to the alt-exchange when a queue is deleted. Additionally, for auto-delete queues with an alt-exchange, upon loss of a primary, the backup brokers should not immediately start the auto-delete process. Instead, the backups should wait for the promotion of a new primary. Upon selection of the new primary, the auto-delete process can start so that the new primary will route the messages to the alt-exchange and the backups will not route the messages to the alt-exchange. Version-Release number of selected component (if applicable): Qpid 0.18 How reproducible: Backup needs to receive the QMF queue delete event prior to receiving the queue's deque replication events. Steps to reproduce: 1) Bind queue test0 to amq.fanout 2) Create queue test1 with an alt-exchange to amq.fanout 3) Send a message to test1 4) Delete queue test1 Actual results: The queue test0 will have 1 message on the primary host but 2 messages on the backups Expected results: The queue test0 will have 1 message on all hosts -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Created] (QPID-5405) [JMS Client] QPid AMQP JMS client does not appear to check for expired messages that have been prefetched.
Timothy Bish created QPID-5405: -- Summary: [JMS Client] QPid AMQP JMS client does not appear to check for expired messages that have been prefetched. Key: QPID-5405 URL: https://issues.apache.org/jira/browse/QPID-5405 Project: Qpid Issue Type: Bug Components: Java Client Affects Versions: 0.24, 0.26 Reporter: Timothy Bish The JMS client does not seem to check for message expiration for messages that have been received and stored in the prefetch buffer and later processed by a synchronous consumer, or a slow async consumer. The Broker used for testing was an ActiveMQ v5.10-SNAPSHOT. The following test shows the basic idea. {code} @Test(timeout=3) public void testTTL() throws Exception { Connection connection = null; try { QueueImpl queue = new QueueImpl(queue:// + name); connection = createConnection(); Session session = connection.createSession(false, Session.AUTO_ACKNOWLEDGE); connection.start(); MessageProducer producer = session.createProducer(queue); producer.setTimeToLive(1000); Message toSend = session.createTextMessage(Sample text); producer.send(toSend); MessageConsumer consumer = session.createConsumer(queue); Message received = consumer.receive(5000); assertNotNull(received); LOG.info(Message JMSExpiration = {}, received.getJMSExpiration()); producer.setTimeToLive(100); producer.send(toSend); TimeUnit.SECONDS.sleep(2); received = consumer.receive(5000); if (received != null) { LOG.info(Message JMSExpiration = {} JMSTimeStamp = {} TTL = {}, new Object[] { received.getJMSExpiration(), received.getJMSTimestamp(), received.getJMSExpiration() - received.getJMSTimestamp()}); } assertNull(received); } finally { connection.close(); } } {code} Adding in some debug in the MessageConsumerImpl show the dispatch of expired messages. {code} private MessageImpl receiveImpl(long timeout) throws JMSException { org.apache.qpid.amqp_1_0.client.Message msg; boolean redelivery; if(_replaymessages.isEmpty()) { checkReceiverError(); msg = receive0(timeout); redelivery = false; } else { msg = _replaymessages.remove(0); redelivery = true; } if(msg != null) { if (isExpired(msg)) { System.out.println(Received expired message:); } preReceiveAction(msg); } return createJMSMessage(msg, redelivery); } private boolean isExpired(org.apache.qpid.amqp_1_0.client.Message msg) { UnsignedInteger ttl = msg.getHeader().getTtl(); Date timeStamp = msg.getProperties().getCreationTime(); if (ttl != null timeStamp != null) { long timeNow = System.currentTimeMillis(); if (timeNow timeStamp.getTime() + ttl.longValue()) { return true; } } return false; } {code} -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Updated] (QPID-5404) HA broker message duplication when deleting a queue with an alt-exchange
[ https://issues.apache.org/jira/browse/QPID-5404?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Alan Conway updated QPID-5404: -- Description: See also: https://bugzilla.redhat.com/show_bug.cgi?id=859119 Description of problem: When deleting a queue with an alternate exchange, both the primary and the backup brokers will route messages to the alt-exchange when the queue is deleted. This will cause the backups to potentially receive duplicate copies of the message if the alt-exchange is bound to a replicated queue (one set from the local route to the alt-exchange and another from the primary due to its route to the alt-exchange). Suggest that the backup brokers should not route messages to the alt-exchange when a queue is deleted. Additionally, for auto-delete queues with an alt-exchange, upon loss of a primary, the backup brokers should not immediately start the auto-delete process. Instead, the backups should wait for the promotion of a new primary. Upon selection of the new primary, the auto-delete process can start so that the new primary will route the messages to the alt-exchange and the backups will not route the messages to the alt-exchange. Version-Release number of selected component (if applicable): Qpid 0.18 How reproducible: Backup needs to receive the QMF queue delete event prior to receiving the queue's deque replication events. Steps to reproduce: 1) Bind queue test0 to amq.fanout 2) Create queue test1 with an alt-exchange to amq.fanout 3) Send a message to test1 4) Delete queue test1 Actual results: The queue test0 will have 1 message on the primary host but 2 messages on the backups Expected results: The queue test0 will have 1 message on all hosts was: Description of problem: When deleting a queue with an alternate exchange, both the primary and the backup brokers will route messages to the alt-exchange when the queue is deleted. This will cause the backups to potentially receive duplicate copies of the message if the alt-exchange is bound to a replicated queue (one set from the local route to the alt-exchange and another from the primary due to its route to the alt-exchange). Suggest that the backup brokers should not route messages to the alt-exchange when a queue is deleted. Additionally, for auto-delete queues with an alt-exchange, upon loss of a primary, the backup brokers should not immediately start the auto-delete process. Instead, the backups should wait for the promotion of a new primary. Upon selection of the new primary, the auto-delete process can start so that the new primary will route the messages to the alt-exchange and the backups will not route the messages to the alt-exchange. Version-Release number of selected component (if applicable): Qpid 0.18 How reproducible: Backup needs to receive the QMF queue delete event prior to receiving the queue's deque replication events. Steps to reproduce: 1) Bind queue test0 to amq.fanout 2) Create queue test1 with an alt-exchange to amq.fanout 3) Send a message to test1 4) Delete queue test1 Actual results: The queue test0 will have 1 message on the primary host but 2 messages on the backups Expected results: The queue test0 will have 1 message on all hosts HA broker message duplication when deleting a queue with an alt-exchange Key: QPID-5404 URL: https://issues.apache.org/jira/browse/QPID-5404 Project: Qpid Issue Type: Bug Components: C++ Clustering Affects Versions: 0.24 Reporter: Alan Conway Assignee: Alan Conway See also: https://bugzilla.redhat.com/show_bug.cgi?id=859119 Description of problem: When deleting a queue with an alternate exchange, both the primary and the backup brokers will route messages to the alt-exchange when the queue is deleted. This will cause the backups to potentially receive duplicate copies of the message if the alt-exchange is bound to a replicated queue (one set from the local route to the alt-exchange and another from the primary due to its route to the alt-exchange). Suggest that the backup brokers should not route messages to the alt-exchange when a queue is deleted. Additionally, for auto-delete queues with an alt-exchange, upon loss of a primary, the backup brokers should not immediately start the auto-delete process. Instead, the backups should wait for the promotion of a new primary. Upon selection of the new primary, the auto-delete process can start so that the new primary will route the messages to the alt-exchange and the backups will not route the messages to the alt-exchange. Version-Release number of selected component (if applicable): Qpid 0.18 How reproducible: Backup needs to receive the QMF queue delete event prior to receiving the queue's deque replication
[jira] [Updated] (QPID-5397) Dispatch - Crash occurs when linked deliveries are concurrently settled
[ https://issues.apache.org/jira/browse/QPID-5397?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ted Ross updated QPID-5397: --- Priority: Blocker (was: Major) Dispatch - Crash occurs when linked deliveries are concurrently settled --- Key: QPID-5397 URL: https://issues.apache.org/jira/browse/QPID-5397 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross Priority: Blocker Dispatch router crashed while running a test in which a client/server pair exchanges requests and responses with acknowledgement (not pre-settled). The test does not throttle the message rates. After several tens-of-thousands of transfers, the router crashed while settling a delivery and attempting to unlink it from its peer. Apparently two linked deliveries were concurrently settled on separate threads. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Resolved] (QPID-5397) Dispatch - Crash occurs when linked deliveries are concurrently settled
[ https://issues.apache.org/jira/browse/QPID-5397?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ted Ross resolved QPID-5397. Resolution: Fixed Dispatch - Crash occurs when linked deliveries are concurrently settled --- Key: QPID-5397 URL: https://issues.apache.org/jira/browse/QPID-5397 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross Priority: Blocker Dispatch router crashed while running a test in which a client/server pair exchanges requests and responses with acknowledgement (not pre-settled). The test does not throttle the message rates. After several tens-of-thousands of transfers, the router crashed while settling a delivery and attempting to unlink it from its peer. Apparently two linked deliveries were concurrently settled on separate threads. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org
[jira] [Commented] (QPID-5397) Dispatch - Crash occurs when linked deliveries are concurrently settled
[ https://issues.apache.org/jira/browse/QPID-5397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13841813#comment-13841813 ] ASF subversion and git services commented on QPID-5397: --- Commit 1548737 from [~tedross] in branch 'dispatch/trunk' [ https://svn.apache.org/r1548737 ] QPID-5397 - Redesigned the fix to the mutex problem - Added a test to detect the settlement-collision problem - Increased the verbosity of the system tests in qdtest - Some minor cleanup in the python router and qdstat Dispatch - Crash occurs when linked deliveries are concurrently settled --- Key: QPID-5397 URL: https://issues.apache.org/jira/browse/QPID-5397 Project: Qpid Issue Type: Bug Components: Qpid Dispatch Reporter: Ted Ross Assignee: Ted Ross Priority: Blocker Dispatch router crashed while running a test in which a client/server pair exchanges requests and responses with acknowledgement (not pre-settled). The test does not throttle the message rates. After several tens-of-thousands of transfers, the router crashed while settling a delivery and attempting to unlink it from its peer. Apparently two linked deliveries were concurrently settled on separate threads. -- This message was sent by Atlassian JIRA (v6.1#6144) - To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org