[EnMasse] Topics and subscriptions in Enmasse 0.24.1

Fri Jan 18 17:26:43 UTC 2019

On Fri, 18 Jan 2019 at 17:27, Bob Claerhout <bob.claerhout at aloxy.io> wrote:

> Hi Rob,
>
> Thanks for your quick response. The sharded topic is indeed working as
> expected. I also saw a new pull request containing a fix for this issue.
> Nice!
>

No problem - sorry you ran into that issue!

> Glad to hear my assumptions(please correct me if I'm wrong) are correct
> concerning the topic (except for the the pattern). Do you have any idea on
> when this pattern matching for subscriptions will be implemented?
>
>
It's not currently scheduled - though we are about to start thinking about
what we'll be working on for the next release after 0.26.  As above, I'm
not saying it definitely doesn't do already do the wildcard matching, I
just don't *think* it does :)  (If you test it and it does indeed work as
expected then let me know and I'll correct my understanding).  It's worth
creating an issue (enhancement) for it in GitHub and then we'll not lose
track of it.

Cheers,
Rob

> Kind regards,
> Bob
>
>
>

> On 1/18/19 5:05 PM, Rob Godfrey wrote:
>
> Hi Bob,
>
> On Fri, 18 Jan 2019 at 15:14, Bob Claerhout <bob.claerhout at aloxy.io>
> wrote:
>
>> Hi all,
>>
>> I'm in the dark about the topics and subscriptions in Enmasse.
>> I'm currently using a standard addressspace in Enmasse 0.24.1 and using
>> the documentation from
>> http://enmasse.io/documentation/0.24.1/openshift/#con-standard-address-space-messaging.
>> Also I'm using qpid-bow as the consumer and publisher (
>> https://github.com/Bynder/qpid-bow).
>> I'm having two questions:
>> My first questions is about the wildcards:
>>
>> The documentation states that:
>>
>>    - a/#/b matches a/foo/b, a/bar/b, and a/foo/bar/b
>>    - a/+/b matches a/foo/b and a/bar/b, but would not match a/foo/bar
>>
>> I've created a topic address "addressspace/test".
>> I'm performing my subscriptions using following command: "qb message
>> receive addressspace/test".
>> I'm performing my publications using following command: "qb message send
>> addressspace/test message".
>>
>> When I'm subscribing to "addressspace/test" and publishing to
>> "addressspace/test" the subscriber receives the message.
>> When I'm subscribing to "addressspace/test/foo" and publishing to
>> "addressspace/test/foo" the subscriber receives the message.
>> However, when I'm subscribing to "addressspace/test/+" and publishing to
>> "addressspace/test/foo" the subscriber does not receive the message
>> although stated by the documentation. When I publish to
>> "addressspace/test/+" it is received by the subscriber.
>> Also, when I'm subscribing to "addressspace/test/#" and publishing to
>> "addressspace/test/foo" the subscriber does not receive the message
>> although stated by the documentation. When I publish to
>> "addressspace/test/#" it is received by the subscriber.
>> When I'm subscribing to "addressspace/test.*" and publishing to
>> "addressspace/test.foo" the subscriber also receives the message. I got
>> this from the automated tests.
>> Can somebody point me towards what I'm doing wrong or did I wrongly
>> interpreted the documentation?
>>
>>
>>
> Unfortunately you are possibly doing nothing wrong, and it is probably a
> bug.  For your topic, which "plan" are you using?  Reviewing the
> configuration it looks like there's a but in EnMasse that for small
> (pooled) topics the pattern matching is not being correctly applied.  For
> large (sharded) topics the pattern matching configuration looks correct in
> the code.  If you are using a pooled topic, can you re-test with a sharded
> topic and see if it works for you then?
>
>
>> My second question is about the subscription address.
>>
>> Just verifying my assumptions:
>> Say I have a topic "test" and I create a subscription, called "foo", on
>> topic "test/foo/+" every message on topic "test/foo/[anything]"(if my
>> assumption in the previous question is correct) should be forwarded and
>> stored in the subscription "foo". When a client connects to "foo", it
>> receives all messages in the queue.
>> When I have the same topic "test" and I create an additional
>> subscription, called "bar", on topic "test/foo/bar" every message on topic
>> "test/foo/bar" should be forwarded and stored in both subscription "foo"
>> and "bar". When a client connects to "foo", it will receive all messages to
>> "test/foo/+" including the ones to "test/foo/bar". When a client connects
>> to "bar" it will receive all message to "test/foo/bar".
>> Are my assumptions correct?
>>
>> Additional sub question:
>> I'm trying to create a subscription using following yaml template, is
>> this correct?
>> - kind: Address
>>   apiVersion: enmasse.io/v1alpha1
>>   metadata:
>>     name: foo
>>   spec:
>>     address: foo
>>     topic: test/foo/+
>>     type: topic
>>     plan: standard-small-topic
>>
>>
>>
> I don't believe that subscription addresses currently support subscribing
> to a pattern (rather only to the root of a topic).  I may be wrong, but
> that is probably an (obvious) enhancement request.
>
>
>> Can somebody please clarify these things for me and maybe provide working
>> examples?
>>
>
> Hope this helps a little
> -- Rob
>
>
>> Thanks!
>>
>> Bob Claerhout
>> --
>>
>> *Bob Claerhout*
>> Software developer
>> M +32 479 34 84 92
>> The Beacon, Sint-Pietersvliet 7, 2000
>> Antwerp
>> bob.claerhout at aloxy.io | www.aloxy.io
>> _______________________________________________
>> enmasse mailing list
>> enmasse at redhat.com
>> https://www.redhat.com/mailman/listinfo/enmasse
>>
>
>
> --
> ____________________________________________________________
> _________________
>
> Red Hat GmbH, www.de.redhat.com,
> Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB
> 153243,
> Managing Directors: Paul Argiry, Charles Cachera, Michael Cunningham,
> Michael O'Neill
>
>
>

-- 
____________________________________________________________
_________________

Red Hat GmbH, www.de.redhat.com,
Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB
153243,
Managing Directors: Paul Argiry, Charles Cachera, Michael Cunningham,
Michael O'Neill
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/enmasse/attachments/20190118/2ea453d5/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: hjgfipjlmcehaiam.png
Type: image/png
Size: 6330 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/enmasse/attachments/20190118/2ea453d5/attachment.png>