launchpad-dev team mailing list archive

[Jonathan Lange <jml@xxxxxxxxxxxxx>]

On Wed, Mar 24, 2010 at 4:17 PM, Cody A.W. Somerville
<cody.somerville@xxxxxxxxxxxxx> wrote:
> Hello,
>
>  I know this has been discussed before but I want to bring it up again
> because I think its important: messaging/task queue system. Just imagine a
> world where doing something on launchpad meant it happened almost instantly
> instead of waiting for the next cronjob to fire! :)

Yes, this would be very good.

> Imagine a world where
> you didn't have to write a schema patches for every little type of job you
> wanted to create but instead just had to write some good old' python code!

Some of us live in this world already.

> Imagine a world where we didn't need our LOASs to be running scripts on
> production servers because we wanted a task or job to occur

I question the relevance of this to message queues.

>. Imagine a world
> where different launchpad applications could easily pass messages and
> communicate with each other without requiring the shackles of strong
> coupling.

You can have strong coupling with a message queue system. Probably
less strong than the mess we have now

This is way more than just having a message queue. Our tightly coupled
mess isn't going to go away by waving a magic technology at it. If we
want to solve that problem, we need a more detailed vision than "bunch
of processes that use a message bus to talk to each other" and we need
concrete steps to get there.

> Wouldn't all that be pretty nice?
>
>  Does anybody who is familiar with the original initiative have any
> information on why its became stalled? What can we do to get this moving
> again?
>

Gary answered this bit. To get it moving again, we need to clear up
other more important stuff in the foundation teams backlog (making the
main webapp faster, getting LP running on lucid in production, using
the Canonical AMIs... see their Kanban board).

jml