Scott,
Currently there isn't a way to do anything that is persistent beyond a
single job. The problem that I ran into with this sort of thing is that if
you create a socket to an external resource in the parent, there is a
possibility that the child processes will try and use the same socket.
However, I am working on some job hooks that would allow you to have some
code run before a job, after a job, before fork, and after fork. Hopefully
these will allow for more cases
where external resources are needed. One thing that you might be able to do
is something like this:
conn = create_connection()
class Job(object)
def perform(*args):
con.do_something()
That might work for each child in the horde process. It should create the
connection the first time it tries to import the job.
Matt
On 2010-08-24, at 10:05 AM, matt george wrote:
> However, I am working on some job hooks that would allow you to have
some code run before a job, after a job, before fork, and after fork.
I think that would work for my situation and would be more generally useful.
Thanks,
scott.