Hello,
Sure when you have time JB let's chat.
Last status : since that's working ok with Spring JmsTemplate, that's
probably the way we shutdown the consumers that is not the same. Still
checking to spot the difference (when I have some time).
Best,
Fred
Le lun. 25 mars 2024 à 10:19,
Hi Fred
Sorry I'm late on that. I'm back there to help you.
The closing looks weird to me. Remember that Session should be on a
single thread (the Connection can be shared across multiple threads,
but the Session should be squared to a single thread).
I suspect something weird around that.
I
Thanks Matt,
Let's continue on github, and let's use the mailing list for a recap later.
Best,
Fred
Le mer. 20 mars 2024 à 16:21, Matt Pavlovich a écrit :
> Hi Frédéric-
>
> A sample project is a big help. First step is to tidy up the dependencies
> and align with Jakarta-based namespace. Using
Hi Frédéric-
A sample project is a big help. First step is to tidy up the dependencies and
align with Jakarta-based namespace. Using ActiveMQ 6.1.0 instead is the best
next step.
I opened a GH issue on your poc repo where other discussions could happen with
less async conversation between
Hello,
Here is a poc : https://github.com/fcurvat/amqpoc
Just start an activemq broker (5.18.3) run the class 'PocAmqApplication'
spring boot app and stop it gracefully (app generate 1000 messages at start
/ got 5 consumers with prefetch 100).
Class 'ShutdownEventSource' allow to hook and stop
Hi,
I will work on setting up a simple project, for now that's tangled with our
code.
Fred
Le mar. 19 mars 2024 à 17:09, Matt Pavlovich a écrit :
> Hi Frédéric-
>
> Do you have a small sample project that is able to reproduce the issue
> that you can share (preferably a simple GitHub
Hi Frédéric-
Do you have a small sample project that is able to reproduce the issue that you
can share (preferably a simple GitHub project)?
Thanks,
Matt Pavlovich
> On Mar 19, 2024, at 10:47 AM, Frédéric Curvat wrote:
>
> Hello,
>
> Checking back on the case, i played with shutdown of
Hello,
Checking back on the case, i played with shutdown of consumers (and not
using brave instrumentation).
With 5 consumers reading, if i close consumers, then sessions, then
connection, i almost always only ack on 4 messages despite 5 messages are
read.
I almost always get one error of this
Hello,
More news about our issue.
We did check again the case and i have some news :
- Nothing bad in broker logs (no poison ack).
- Application logs shows that issue appears on graceful shutdown of the
application (however not on all shutdowns)
Our shutdown consists in calling .close() method
Hello JB !
Hope you are well !
>1. The message goes in redelivery (because it expired or client
>rollback transaction) and so it can be taken by another consumer. As
>you use session_transacted, the "first" client has to deal with the
>rollback
For me that was not so likely because the message
Hi Matt !
Thanks for your answer. I understand email support for this kind of issue
is a bit hard.
We will double check on all this, check all logs.
I also understand that logging the messageId would be a great help to
identify what is happening, so we might consider that.
Best,
Fred
Le ven. 1
Oh by the way, what's your consumer prefetch ? I guess it's more than 1 ?
Regards
JB
On Fri, Mar 1, 2024 at 4:52 PM Frédéric Curvat wrote:
>
> Hello !
>
> At my company we are using Apache ActiveMQ 5.18.3.
> We suspect that in some rare cases, a queue message is read twice by
> different
Hi Fred
I see only two options:
1. The message goes in redelivery (because it expired or client
rollback transaction) and so it can be taken by another consumer. As
you use session_transacted, the "first" client has to deal with the
rollback
2. Do you see "poison ack" in the log ?
Feel free to
Hi Fred-
Most likely, you are seeing a _redelivery_ vs a bug or actual duplicate
message.
First thing to confirm that it is in fact the same message by comparing
messageId, and not just properties or message body content. There are scenarios
where a producer may inadvertently send the same
Hello !
At my company we are using Apache ActiveMQ 5.18.3.
We suspect that in some rare cases, a queue message is read twice by
different consumers.
For more context :
- broker is classic primary/secondary (secondary started but not active -
not a network of brokers).
- we are using persisted
15 matches
Mail list logo