OpenERP cronjobs not always work... It would be very nice if we can use the queue system to do some cronjobs, like celery does.

The idea is make a decorator cron like:

@cron(min, hour, dom, month, dow)
def some_cron_function(self, cursor, uid, context=None):
    pass

Then we need to register cronjob function. Find a way to do it elegantly..

That's all.

AUTOWORKER_REDIS_URL is forced with redis:// -> rediss:// for ssl connections?
self.connection = Redis.from_url(self.config['redis_url']) # this is called on autoworker init and if

def setup_redis_connection():
    redis_conn = get_current_connection()
    if not redis_conn:
        if config.get('redis_url', False):
            oorq_log('Connecting to redis using redis_url: %s'
                     % config['redis_url'])
            redis_conn = from_url(config['redis_url'])
        else:
            oorq_log('Connecting to redis using defaults')
            redis_conn = Redis()
        push_connection(redis_conn)
    os.environ['AUTOWORKER_REDIS_URL'] = get_redis_url(redis_conn)
    return redis_conn


def get_redis_url(redis_conn):
    """
    Creates redis url from redis connection
    :param redis_conn:
    :return: url on the form redis://host:port/db
    """
    if not redis_conn:
        return False
    return 'redis://{host}:{port}/{db}'.format(
        **redis_conn.connection_pool.connection_kwargs
    )

We should review change introduced in rq/rq#1182 because it cleans automatically failed jobs, and our script to requeue jobs will fail.

https://github.com/nvie/rq/blob/master/rq/registry.py

requirements.txt is a symbolic link and that is a problem for windows. It downloads it like a normal text file with this contents:

../requirements.txt

When updating this module or updating all modules pip fails installs to process the file.

Can it be changed to a normal text file? If you hate repeating your self, may be the root file can be the symbolic link to the one inside oorq folder.

Maybe we can use a special worker script to preload OpenObject libraries and then speed up the job process.

There is an example in Performance notes http://python-rq.org/docs/workers/

If we detect that OPENERP_SENTRY exists we have to change the transport to not be asynchronous

@ecarreras ,

I've just come across this module on Github. My apologies for writing a question in the form of a github issue I was just not sure how to contact you.

Regarding this module, would this be suitable for setting up as a caching mechanism in Odoo 8.0? I'm working on implementing a module (product_multi_image - https://github.com/pedrobaeza/product-attribute/tree/8.0-product_images WIP) to store and display multiple product images on the backend (sales >> product), and front-end (website_sale product pages).

Due to the size of the catalog i'll be building I think it may be necessary for me to implement a caching and cache-warming module in order to keep the instances running with decent performance.

I've heard of Redis being used with Odoo in various forum posts, however there is not much information publicly available on the subject.

In api_v5 we have a Worker Class that speeds up worker jobs.

Pros

• Speed up time between jobs.

Cons

• We have to restart the worker to reload the code.