I am currently playing around with Dancer2 and wonder about the proper way to start a long-running child-process.
But even though I currently use Dancer2 I guess my question is more about PSGI in general.
I don't know that much about it but my understanding is that PSGI/Plack abstracts away the deployment details, so in theory your PSGI application should not be affected regardless of whether the app is later deployed as GCI or running under starman or whatever, but I cannot see how you can abstract away all the mechanics of forking (e.g. what part of the system reaps the childs again etc).
So what is the best practice here?
Not forking at all and using some job-qeueue mechanism or dealing with it on a case-by-case basis as there is no abstraction for it?
Please enlighten me...
In reply to How to fork in PSGI/Dancer2 by morgon
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |