You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by IlarioM <im...@gmail.com> on 2017/04/14 16:04:23 UTC
Artemis subscription timing
hi,
i've written a simple junit test where an mqtt client subscribe to 50000
different topics to check how many times is needed;
i log timing every 50 subcriptions at the beginning everythings run fine
(0.1sec for 50 topics) but after a while performance start degrading and
after 10 mins
i meausure 1sec for 50 topics(if i run same tests on a clean artemis
installation i see same behaviour but delays are more than doubled)
Are these delays optimizable? Am I missing something?
PahoMQTTSubscriberTest.java
<http://activemq.2283324.n4.nabble.com/file/n4724964/PahoMQTTSubscriberTest.java>
I'm using a centos virtual machine inside vbox with i5 processor and 4gb of
ram (8 gb total ram )
Thank you very much
--
View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Artemis subscription timing
Posted by Clebert Suconic <cl...@gmail.com>.
When you create a queue on a messaging system, you are requesting
guarantees of delivery, even if you disconnect consumers.. etc.. so
the queues will be created on server.. and message references will be
created for the consumer you requested...
So, ou can't just keep creating subscriptions on a topic.. you create
and reuse them..
On Wed, Apr 19, 2017 at 6:47 AM, IlarioM <im...@gmail.com> wrote:
> Hi Clebert,
> thanks for your response; I just runned the same test with jms clients and i
> see the same behaviour
>
> TopicExample2.java
> <http://activemq.2283324.n4.nabble.com/file/n4725076/TopicExample2.txt>
>
>
> please can you give me more info about these delays? There is a way to
> extimate delay for an incoming subscription?
>
>
> Thank you very much
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964p4725076.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
--
Clebert Suconic
Re: Artemis subscription timing
Posted by IlarioM <im...@gmail.com>.
Hi Clebert,
thanks for your response; I just runned the same test with jms clients and i
see the same behaviour
TopicExample2.java
<http://activemq.2283324.n4.nabble.com/file/n4725076/TopicExample2.txt>
please can you give me more info about these delays? There is a way to
extimate delay for an incoming subscription?
Thank you very much
--
View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964p4725076.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Artemis subscription timing
Posted by Clebert Suconic <cl...@gmail.com>.
I know Martyn had a change about MQTt and what u are dealing with on
master.
we will release 2.1 very soon. I even sent a heads up on the dev list but I
have a personal conflict.
On Sat, Apr 15, 2017 at 9:36 AM IlarioM <im...@gmail.com> wrote:
> Hi Clebert,
> I'm using Artemis 2.0.0
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964p4724973.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
--
Clebert Suconic
Re: Artemis subscription timing
Posted by IlarioM <im...@gmail.com>.
Hi Clebert,
I'm using Artemis 2.0.0
--
View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964p4724973.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Artemis subscription timing
Posted by Clebert Suconic <cl...@gmail.com>.
What version ?
On Fri, Apr 14, 2017 at 4:35 PM IlarioM <im...@gmail.com> wrote:
> hi,
> i've written a simple junit test where an mqtt client subscribe to 50000
> different topics to check how many times is needed;
> i log timing every 50 subcriptions at the beginning everythings run fine
> (0.1sec for 50 topics) but after a while performance start degrading and
> after 10 mins
> i meausure 1sec for 50 topics(if i run same tests on a clean artemis
> installation i see same behaviour but delays are more than doubled)
> Are these delays optimizable? Am I missing something?
>
> PahoMQTTSubscriberTest.java
> <
> http://activemq.2283324.n4.nabble.com/file/n4724964/PahoMQTTSubscriberTest.java
> >
>
> I'm using a centos virtual machine inside vbox with i5 processor and 4gb of
> ram (8 gb total ram )
>
> Thank you very much
>
>
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Artemis-subscription-timing-tp4724964.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
--
Clebert Suconic