[wdmmg-dev] Where we are with OpenSpending

[Martin Keegan]

>
> The more serious issue is that we are having trouble with the task queue,
> "celery". An option has been added to bypass celery, but this only really
> works for small datasets; users trying to load large datasets will get a
> timeout.
>
>
> I originally setup celery for openspending. As you noticed yourself,
> asynchronous processing is necessary and circumventing it can't possibly be
> a viable solution.
> Can you describe the problem you are having with celery?
>
> Indeed: circumventing it isn't viable for most cases.

The problem we're having is that celery seems to get into a state where it
stops processing new tasks.

It looks like the tools for checking celery's status (cli or python) don't
work terribly well: ergo we can't pass this information on to end users.

Mk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/openspending-dev/attachments/20110615/d2b800cb/attachment.html>