@ejsmont-artur

Sorry for a delayed reply.

I would personally stay with cron and use script like this one I have created recently:

https://github.com/laelaps/symfony-gearman-bundle/blob/master/Tool/php-watcher.sh

If you run this script every minute with cron (I think this is a reasonable time) it will:

• check if there is a new worker matching given pattern and if so - run it
• check if every worker is running and if not - run it
• check if source code of worker changed after running it (!) and if so - restart this one particular worker

Sample usage is:

$ php-watcher.sh "/full/path/to/your/workers/*php" "/full/path/to/your/log/directory" "php /full/path/to/app/console gearman:worker:run"

This way you can be sure that even in an emergency situation (for example - long running worker crashed after exceeding memory limit), not a single worker will be down for longer than one minute.

This way you can safely order script to kill itself with exit or die anytime and be sure that it will return in a minute.

Thanks man, in the mean time i found supervisors, its quite awesome so ill go with it.

http://supervisord.org/

Cheers

Art