launchpad-dev team mailing list archive

[Robert Collins <robert.collins@xxxxxxxxxxxxx>]

On Wed, Jun 22, 2011 at 4:07 AM, Gary Poster <gary.poster@xxxxxxxxxxxxx> wrote:
>
> On Jun 21, 2011, at 9:14 AM, Robert Collins wrote:
>
>> So we kindof got into an analysis paralysis situation with Rabbit and
>> persistence. Specifically we don't know how much HA we need, or when
>> we need it. This makes it hard to say that rabbit is/isn't HA enough
>> for us.
>>
>> http://zguide.zeromq.org/page:all is well worth reading - irrespective
>> of our choices, its got a very well reasoned layout of many issues and
>> antipatterns that can happen.
>
> I enjoyed what I read so far of the zeromq page.  I'll hope to come back to it later.  The oddest thing I saw so far was the consumer PULL effectively taking all available messages from a producer PUSH, as fas as I could tell, rather than being able to specify to only take "as much as you can chew right now" or something.  Maybe that's addressed later in the document.
>
> Has the Rabbit-on-top-of-0MQ code been mentioned here?  I haven't seen it discussed, but I might have missed it.
>
> https://github.com/rabbitmq/rmq-0mq/wiki/
>
> The "Known limitations" discussed at the bottom of that page look serious, but it also looks like this is early days for the integration project.

Its a pretty cool thing; Elliot pointed it at me the other day.

It may make sense as a bridge in future, or we may want to migrate
over stacks entirely (but I couldn't argue that 0mq is what we need
clearly and unambiguously today :)). As a project I think we're still
getting our heads around eventual consistency; thats a pretty core
change on its own, and until its sunk in we'll find it hard to really
reason about things.

-Rob