I've made some further changes to resolve an issue that might have
been something you hit in your scenario, where if you have the JMS 1.1
and JMS2 APIs both on the classpath it could fail during send, so I've
tweaked the client to tolerate that situation. Regarding the
thread/heap usage, thats
Unfortunately the latest update is not working for me either. Though it
seems that connection is established to the IoT hub, the messages cannot be
delivered. If I look at the AMQPS frames, I can see only Begin and Empty
frames being exchanged.
Also AmqpProvider and QpidJMS Connection Executor
I've made some changes to handle sending foreign messages from JMS 1.1
providers and published a new snapshot, you should be able to give
that a try now.
Robbie
On 13 December 2016 at 10:47, Robbie Gemmell wrote:
> Yes, that appears to be the client trying to call
Yes, that appears to be the client trying to call setJMSDeliveryTime
on the Message, only to find the Message object impl doesnt support
that due to being from a JMS 1.1 impl. Will need to have a think about
how best to handle that. As a workaround currently you'd need to
either modify the client
Seems to me that Qpid JMS client is not compatible with my setup. I tried
to use latest snapshot as dependency, but at the moment I'm getting:
2016-12-13 07:59:11,815 | ERROR | Error in thread 'ActiveMQ Session Task-1'
| org.apache.activemq.thread.TaskRunnerFactory | ActiveMQ Session Task-1
Tomorow I plan to do some testing with 0.20.0-SNAPSHOT. The thing is I can
hardly use snapshots in my company's software (next release version of the
OSGi bundle I'm currently working on). So even if the fix made in
QPIDJMS-228 solves all my problems I have to wait for a release or find a
Thanks! I'll try that as soon as 0.20.0 is released!
This might also help with another issue I'm experiencing - The
ThreadPoolExecutor in JmsConnection is producing non-daemon threads
intentionally (I can see the comment, but still don't get the reason why to
do it in your code.) Which means in
Close timeout handling has been fixed up a fair bit with issue:
https://issues.apache.org/jira/browse/QPIDJMS-228
On 11/25/2016 11:04 AM, Robbie Gemmell wrote:
"I also think I see an issue with the closeTimeout handling in that in
one place it seems likely to be using the requestTimeout value
"I also think I see an issue with the closeTimeout handling in that in
one place it seems likely to be using the requestTimeout value
instead. Need to investigate that further."
Though in my attemps that didnt actually matter and the close
completed fine if the network dropped and no response
I'll try to find a way of preventing this problem from happening. (Though I'm
not sure how to do that properly at the moment.)
But how do you explain unlimited waiting for Close frame response as shown
in my previous post?
DraCzech wrote
> 2016-11-24 16:41:45.202589 [19105288:0] ->
I had a look at this on master and though I did see a missing guard
(diff below) on the connection close that could be related, I haven't
reproduced what you are seeing even without that. It may still be
worth you giving it a try out.
Your logging snippets only include protons protocol trace, it
Yes that looks like its trying to get a lock which the onMessage
handler is holding while firing, which is ongoing due to a send within
it.
On 25 November 2016 at 07:37, DraCzech wrote:
> When going through logs I also found this piece of thread dump. Which shows
> that my
When going through logs I also found this piece of thread dump. Which shows
that my cleaning thread is waiting for a lock held by ActiveMQ thread. A
state which is definitely not desired, but I'm not sure if can be prevented
completely...
2016-11-24 17:46:10.016523 "JMS Connector cleaner" daemon
So I tried to turn on logging on qpid and this is what I found so far.
In case I shut down the messaging bundle, it sends *Close*, receives
*Close*, everything goes fine.
But In case the device loses network connection it sends *Close* and waits
for response which never comes.
I'm using
I'm not sure it would relate to those bits, and could well just be a
bug. Can you give a bit more detail around e.g. your uri settings etc.
You might also want to turn up the logging and see if anything stands
out, see
http://qpid.apache.org/releases/qpid-jms-0.11.1/docs/index.html#logging.
Hi, I'm currently using ActiveMQ 5.14.1 embedded broker connected from
several local queues to MS Azure IoT Hub using JmsConnectionFactory as
foreignConnectionFactory in SimpleJmsQueueConnector. (Using dependency to
qpid jms client 0.11.1) Each queue is connected to the server via its
respective
16 matches
Mail list logo