activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gary Tully <gary.tu...@gmail.com>
Subject Re: [DISCUSS} HornetQ & ActiveMQ's next generation
Date Thu, 26 Mar 2015 12:29:44 GMT
Hadrian,

working through your rationale
on 1, there is nothing stopping any member from educating themselves
and making informed decisions.
on 2, the existing committers identify the synergy that is possible
between the code bases. That too comes from education via the code.
on 3, what is the basis for that statement?

To my mind there is a big mix up between a code grant and hijacking a
community. The grant is in good faith and has been accepted in good
faith. 5.x needs a new core. If 6.0.1 or 6.0.2 or whatever it is
called does not fly let 5.x skip 6 and go to 7. The versioning thing
is not a big deal. The reason 6.0 is sensible in my mind is that it
captures intent. It is quite different but it supports openwire, so it
is a drop in replacement. It is java code, xml config, has near
capability parity with the exception of a jdbc store. It is already
production quality code.

There is so much goodness there that the next 6 months can be spent on
incremental improvements to the migration strategy rather than on a
rewrite of 5.x.

I don't know who is suggesting that 5.x will die with or without 6.
5.x has quite a bit of in built inertia that will sustain it.
I do know that 5.x has some limitations and the same limitations do
not exist in the code grant source.

I think a lot of confusion comes from the website (next generation
...) and the lack of incremental communication on the progress of the
code grant.

I would love to have some consensus on direction.


On 26 March 2015 at 00:44, Hadrian Zbarcea <hzbarcea@gmail.com> wrote:
> Sure absolutely.
>
> I changed my mind a few times during this thread. The rationale for my
> revised recommendation is:
>
> 1. Many in the activemq community accept the notion that hornetq has
> technical merits, yet very few actually looked at the code. We are nothing
> more than a peanut gallery for the hornetq guys.
> 2. If I were one of the active committers in hornetq, I would jump at the
> opportunity to build the project and community without the activemq baggage.
> Independently (in the incubator) the project has the freedom to go whatever
> path they choose.
> 3. In the incubator they would benefit from far better mentoring than what
> they get in the ActiveMQ pmc.
>
> Yes, the hornetq commmitters did a lot of work (didn't get much help also).
> It would be frustrating and a bit unfair to them. Yes, their opinion does
> count.
>
> Renaming the subproject to something else than activemq6 was not accepted.
> This would have been imho a very simple solution until such time that a
> strong community exists. By that time, it might be activemq 7 or 8, who
> knows. The strong push from one side of the fence to keep it as activemq6,
> concerns me. It sounds like the hornetq proponents feel that the only way to
> be successful is to (what is perceived as) hijack of the activemq brand and
> future of the project. And the idea that the activemq project will die
> without the infusion of hornetq is, well, insulting. What facts support that
> statement? I, personally, react very strongly to such manipulative comments.
>
> Now that Chris is aware of this threads, I think enough points of view were
> expressed and I hope we'll soon hear a more authoritative opinion.
>
> Best,
> Hadrian
>
>
>
> On 03/25/2015 06:52 PM, Rob Davies wrote:
>>
>> Hadrian  - so the original vote was to accept the code into ActiveMQ
>> specifically. We would have to check with the HornetQ developers to see if
>> now going to incubator instead was acceptable to them - but it does seem to
>> me that their desire to join ActiveMQ was borne out of consolidation of the
>> code bases - and it does appear it was accepted in that vein. The HornetQ
>> code base has been worked on to be ready to pass ASF release guidelines -
>> but actually trying to release it had caused a lot of concern within the
>> community, which is understandable.
>>
>> So instead of passing HornetQ to the incubator - wouldn't it be better to
>> start a discussion on how some/all/none of it  could be incorporated into
>> ActiveMQ? I'm not sure this had happened yet.
>>
>> Rob
>>
>>
>>
>>> On 25 Mar 2015, at 20:53, Hadrian Zbarcea <hzbarcea@gmail.com> wrote:
>>>
>>> Surely, calling it HornetQ (or whatever name the community chooses) and
>>> building the community in the incubator does not prevent anything you
>>> mentioned from happing, right?
>>>
>>> Apollo did it right actually. Some folks said that they had a better
>>> idea, called the project Apollo, not activemq-6 (although there were the
>>> same references to plans for apollo to become the core of the next
>>> generation of activemq). And the Apollo community went on to prove its
>>> viability. The fact that it didn't happen has many reasons, most of them not
>>> technical.
>>>
>>> As an ASF member I am interested in the viability and maturity of a
>>> project community. HornetQ is not yet there. And the fact fact that it's
>>> called activemq-6 effectively prevents the current mature activemq project
>>> to have a version 6 in any shape that is not HornetQ.
>>>
>>> I was kinda neutral initially and did not get involved in this thread
>>> initially. But the passion to keep the activemq6 name for hornetq, makes me
>>> very suspicious (coupled with past experiences) that I am getting very
>>> strongly in favour of hornetq being re-hosted in the incubator where there
>>> are many very experienced ASFer that could mentor and assist (more than in
>>> the ActiveMQ community). If HornetQ will not happen, the same way Apollo
>>> didn't, this will prove to be another distraction. Some are excited about
>>> the future, others are frustrated about the present.
>>>
>>> Cheers,
>>> Hadrian
>>>
>>>
>>>> On 03/25/2015 03:00 PM, Christopher Shannon wrote:
>>>> My team has been using ActiveMQ pretty heavily for the past couple of
>>>> years
>>>> and we have been following the ActiveMQ 6 discussion from the beginning
>>>> last summer.  It was always our impression that HornetQ was going to
>>>> become
>>>> the core of the next broker (just like Apollo originally was) and
>>>> overall
>>>> our team is excited about the idea of going with HornetQ as the next
>>>> generation of ActiveMQ.
>>>>
>>>> Mostly, our reasons are from the technical side of things.  My team is
>>>> in
>>>> the source code every day making modifications for our own needs and
>>>> ActiveMQ certainly is showing its age.  Hiram has already made good
>>>> points
>>>> earlier as to why HornetQ would provide a better foundation from a
>>>> technical standpoint going forward.  Another really big thing for us is
>>>> that our team has been waiting for JMS 2.0 support for a while and it
>>>> doesn't seem like anyone wants to support it in the current code base.
>>>> We
>>>> need to be able to support a variety of protocols (STOMP, JMS 1/2, AMQP,
>>>> etc) which is something HornetQ does.
>>>>
>>>> I think that calling it ActiveMQ 6 is the way to go as long as it still
>>>> supports all of the features in ActiveMQ 5.x (Virtual Destinations,
>>>> OpenWire, etc) before going final and there are clear migration
>>>> instructions.  We would need to have ActiveMQ 6 either support KahaDB or
>>>> there would need to be a way to migrate existing data to the new data
>>>> store
>>>> type.  We would also need to be able to have a network of brokers that
>>>> include both an ActiveMQ 5.x broker and an ActiveMQ 6.x broker.  Lastly,
>>>> it
>>>> would also be nice to have a roadmap posted and kept up to date so we
>>>> can
>>>> track the progress of the code and test out milestone releases.
>>>>
>>>> In my opinion having another sub project (along with Apollo) would just
>>>> make things even more confusing.  As Andy pointed out, having everyone
>>>> in
>>>> the community join together to support one broker going forward would
>>>> produce a better broker than by splitting up resources and potentially
>>>> causing it to die out.
>>>>
>>>>> On Wed, Mar 25, 2015 at 2:43 PM, Andy Taylor <andy.tayls67@gmail.com>
>>>>> wrote:
>>>>>
>>>>> Rather than the activemq community jumping ship and leaving it to sink
>>>>> at
>>>>> some point in the future, let's ensure the future of activemq and its
>>>>> community and actually grow it by bringing 2 communities together by
>>>>> having
>>>>> a project tbat everyone could (and should) get behind.
>>>>> On 25 Mar 2015 18:27, "Hadrian Zbarcea" <hzbarcea@gmail.com> wrote:
>>>>>
>>>>>> This is not a view shared by everybody.
>>>>>>
>>>>>> The way I read Chris' mail is that hornetq should have actually
>>>>>> started
>>>>>
>>>>> in
>>>>>>
>>>>>> the incubator and build a community as the next best messaging
>>>>>> solution.
>>>>>
>>>>> If
>>>>>>
>>>>>> hornetq succeeds, it is possible that some (or all) from the activemq
>>>>>> community will jump boat. Who knows.
>>>>>>
>>>>>> But why undercut the current activemq project? HornetQ can very well
>>>>>> be
>>>>>> the solution you mention in the incubator right?
>>>>>>
>>>>>> After all this long discussion, my recommendation is to move hornetq
>>>>>> in
>>>>>> the incubator and let it evolve over there. It would be beneficial
for
>>>>>
>>>>> for
>>>>>>
>>>>>> the hornetq project too to grow without the activemq distraction.
They
>>>>>
>>>>> can
>>>>>>
>>>>>> choose to be as close or distant they want from the current activemq
>>>>>> features. The activemq community is obviously biased towards what
>>>>>
>>>>> activemq6
>>>>>>
>>>>>> should offer and that may or may not jive with the vision the hornetq
>>>>>> community has for their project.
>>>>>>
>>>>>> Cheers,
>>>>>> Hadrian
>>>>>>
>>>>>>
>>>>>>> On 03/25/2015 01:56 PM, David Jencks wrote:
>>>>>>>
>>>>>>> Sorry, can't stop typing.
>>>>>>>
>>>>>>> My impression is the problem hornetQ is a solution for is that
anyone
>>>>>>> picking a messaging solution based on technical rather than political
>>>>>>> factors is not going to pick activemq.  I thought Hiram said
this
>>>>>>> pretty
>>>>>>> explicitly.  Did I misunderstand?
>>>>>>>
>>>>>>> thanks
>>>>>>> david jencks
>>>>>>>
>>>>>>> On Mar 25, 2015, at 12:05 PM, artnaseef <art@artnaseef.com>
wrote:
>>>>>>>
>>>>>>>   Growing the community around HornetQ is the same issue regardless
>>>>>>> of
>>>>>
>>>>> the
>>>>>>>>
>>>>>>>> naming - it needs to happen, and just naming it ActiveMQ
6 doesn't
>>>>>
>>>>> really
>>>>>>>>
>>>>>>>> change anything other than to create the presumption that
HornetQ
>>>>>>>> will
>>>>>>>> succeed as ActiveMQ 6.
>>>>>>>>
>>>>>>>> Sharing a direction across the community is important, and
making
>>>>>>>> sure
>>>>>>>> that
>>>>>>>> direction is clear is also important.  In that light, I am
very glad
>>>>>>>> to
>>>>>>>> be
>>>>>>>> having this discussion.
>>>>>>>>
>>>>>>>> The statement "Neglecting to commit to a direction will leave
>>>>>>>> ActiveMQ
>>>>>>>> rudderless" is valid, but does not decide that direction.
 Nor does
>>>>>>>> it
>>>>>>>> mean
>>>>>>>> that a complete restart of ActiveMQ is the right direction.
>>>>>>>>
>>>>>>>> So, let's put this back into perspective.
>>>>>>>>
>>>>>>>> We have the HornetQ donation to ActiveMQ.  To what benefit
for the
>>>>>>>> ActiveMQ
>>>>>>>> community?  Age of the solution is not a compelling argument
>>>>>>>> (consider
>>>>>>>> that
>>>>>>>> Java is even older than ActiveMQ).
>>>>>>>>
>>>>>>>> ActiveMQ continues to be very widely used and supported.
 It serves
>>>>>>>> mission-critical functions in large companies across multiple
>>>>>
>>>>> industries,
>>>>>>>>
>>>>>>>> and even supports critical government infrastructure in many
places.
>>>>>>>>
>>>>>>>> Only time will tell if HornetQ is up to the task on all fronts:
>>>>>
>>>>> strength
>>>>>>>>
>>>>>>>> of
>>>>>>>> technology; community to maintain, support, and advocate
the
>>>>>
>>>>> technology;
>>>>>>>>
>>>>>>>> ease of installation, use, and monitoring; etc.  Therefore,
a
>>>>>
>>>>> presumption
>>>>>>>>
>>>>>>>> that it will replace an existing, proven solution is premature.
>>>>>>>>
>>>>>>>> Really, the merits here are hard to argue because I'm not
seeing any
>>>>>>>> valid
>>>>>>>> merits described.
>>>>>>>>
>>>>>>>> I keep wondering, "what problem are we solving?"  Please
help me to
>>>>>>>> understand this and how the HornetQ donation solves the problem.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> View this message in context: http://activemq.2283324.n4.
>>>>>>>> nabble.com/DISCUSS-HornetQ-ActiveMQ-s-next-generation-
>>>>>>>> tp4693781p4693805.html
>>>>>>>> Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
>>>>
>>>>
>

Mime
View raw message