Hi Jayadevan,

I’ve been thinking about writing a better docs article about the various run 
modes for uWSGI, or even just expanding the woefully inadequate glossary we 
have right now to cover this sort of stuff, but for now...

In master (--master) mode, an uWSGI master process governs 1 to N workers 
(--workers/--processes N). Each worker can process a single request at a time 
(when threaded/async modes are not used). (You can also run uWSGI without a 
master, but that’s usually not the case nor a good idea as far as I know.)
The master keeps track of its workers, making sure they don’t misbehave (memory 
limits, request limits, what-have-you) and that enough of them are up at any 
given time (`cheaper`/other autoscaling modes).
A single application would usually run under a single master.

The emperor/vassal system is conceptually one layer higher than this. The 
emperor governs different applications, starting and stopping them according to 
configuration changes. You would use emperor mode if you had several 
sites/applications running on the same server. You don’t necessarily need to 
run emperor in master mode, though it is possible.

Does this help at all? :)

Regards,
Aarni

From: [email protected] [mailto:[email protected]] On 
Behalf Of Jayadevan Maymala
Sent: Wednesday, November 27, 2013 5:01 PM
To: uWSGI developers and users list
Subject: [uWSGI] documentation

Hi,
Is there a blog/doc where the concepts of
Emperor > Vassal
Master > Worker etc are provided? I am trying to figure out how they interact 
with each other, their function etc.
Regards,
Jayadevan
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to