>> I think the easiest way would be having a persistent zergpool (it is an
>> instance running to pass the socket to zergs) and simply suspend/resume
>> the various instances.
>>
>> Zergpools are described at the end of the zerg page (zergpools instances
>> can be managed by the emperor too)
>> Roberto De Ioris
>
>
> regular zergmode with a unix port is working well for me.  I tried
> setting up zergpools in my ini and they do not seem to work.  I.e. if
> I SIGTSTP the main uwsgi master then nothing happens, whereas with
> regular zerg server and zerg master the zergs take all requests.  This
> is on Mac OsX with uWSGI 1.9.6 for development mode where we use http
> = <port> instead of socket = <port>.  in production we are behind
> nginx and use socket mode.  I see the pool created and zerg server
> setup, but it never does any work.  Am I set up/doing something wrong?
>
> thanks
>
>
> uwsgi --ini uwsgi-development.ini
>
> uwsgi --zerg /tmp/zergpool_1 --ini uwsgi-development.ini:zerg
>
> [uwsgi]
> env = PYRAMID_INI=development.ini
> http = :6543
> master = true
> processes = 4
> harakiri = 60
> harakiri-verbose = true
> limit-post = 524288000
> post-buffering = 8192
> listen = 256
> #max-requests = 1000
> #reload-on-as = 128
> reload-on-rss = 96
> no-orphans = true
> log-slow = true
> enable-threads = true
> virtualenv = ../env
> wsgi-file = wsgi.py
> #touch-chain-reload = ../Common/conf/restart
> zergpool = /tmp/zergpool_1:127.0.0.1:6544,127.0.0.1:6545
> #zerg-server = /tmp/zergport
> pidfile = /tmp/uwsgi_master.pid
>
> [zerg]
> env = PYRAMID_INI=development.ini
> #socket = :6544
> master = true
> processes = 4
> harakiri = 60
> harakiri-verbose = true
> limit-post = 524288000
> post-buffering = 8192
> listen = 256
> #max-requests = 1000
> #reload-on-as = 128
> reload-on-rss = 96
> no-orphans = true
> log-slow = true
> enable-threads = true
> virtualenv = ../env
> wsgi-file = wsgi.py
> #zerg-fallback = true
> pidfile = /tmp/uwsgi_zerg.pid
> _______________________________________________
> uWSGI mailing list
> [email protected]
> http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
>

zergpools must be run separately from regular instances, you have a vassal
for the zergpool, one for the first instance and one for the second.

The zergpool vassal is never touched, the others constantly "dance"
taking/leaving the socket

-- 
Roberto De Ioris
http://unbit.it
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to