Uploaded image for project: 'AMQ Broker'
  1. AMQ Broker
  2. ENTMQBR-847

Deadlock detected on large number of messages

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Critical Critical
    • None
    • None
    • broker-core
    • User Experience
    • Hide

      On Fedora 25:

      • git clone https://github.com/apache/qpid-proton
      • cd qpid-proton
      • git checkout 0.18.0-beta
      • mkdir build
      • cd build
      • cmake ..
      • cmake --build .
      • sudo make install
      • cd ../..
      • sudo dnf install librdkafka librdkafka-devel [not active but used in src]
      • git clone https://github.com/RoddieKieley/foranton
      • cd foranton
      • git checkout linux
      • mkdir build
      • cd build
      • cmake ..
      • cmake --build .
      • update amq.properties for broker location and 2 producers and 2 consumer if on 1 box
      • Install broker 7.0.2.SE458
      • disable paging on the broker; disable other protocols except AMQP on 5672
      • run broker
      • run ./test_client amq.properties on each machine or on single machine depending
      Show
      On Fedora 25: git clone https://github.com/apache/qpid-proton cd qpid-proton git checkout 0.18.0-beta mkdir build cd build cmake .. cmake --build . sudo make install cd ../.. sudo dnf install librdkafka librdkafka-devel [not active but used in src] git clone https://github.com/RoddieKieley/foranton cd foranton git checkout linux mkdir build cd build cmake .. cmake --build . update amq.properties for broker location and 2 producers and 2 consumer if on 1 box Install broker 7.0.2.SE458 disable paging on the broker; disable other protocols except AMQP on 5672 run broker run ./test_client amq.properties on each machine or on single machine depending
    • AMQ Broker 7.1 Sprint 2

      When there is a large amount of messages for consumers to read and they are processing them at speed (achievable using c++ client with 2 producers and 2 consumers), we notice the following log on the broker side:

      09:25:25,811 WARN  [org.apache.activemq.artemis.core.server] AMQ222174: Queue example, on address=example, is taking too long to flush deliveries. Watch out for frozen clients.

      Though even taking the consumers off, and having only one the issue remains, it only resolves by bouncing the broker.

      On googling we notice a similar issue https://issues.apache.org/jira/browse/ARTEMIS-450 still open.

      Requires broker bounce to resolve.

              csuconic@redhat.com Clebert Suconic
              rhn-support-rkieley Roderick Kieley
              Andy Taylor Andy Taylor
              Votes:
              0 Vote for this issue
              Watchers:
              10 Start watching this issue

                Created:
                Updated:
                Resolved: