Le lundi 19 septembre 2016 13:34:06 UTC+2, Sergi Almacellas Abellana a écrit : > > El 19/09/16 a les 10:26, Ali Kefia ha escrit: > > We faced this problem last year and worked on scaling trytond this way > > > > * Shared Cache on Redis > > * Load balance with Nginx > > * Sync mechanism via redis PUB/SUB to reload Pool > > > > Description of Coog (trytond spécialized ERP) > > here: http://coopengo.com/coog-v1-6-nouvelle-architecture/ > > Implementation on a trytond fork: https://github.com/coopengo/trytond > > For more details or help using it, send me an email > > AFAIU you need the redis chache and the sync mechaninsm because you have > the wsgi servers running in different machines. Am I right? >
Different processes: this work has been done before werkzeug We tried to investigate ir.cache and did not succeed to ensure propagating invalidation BTW, with Redis, data is not duplicated on memory (when having 10 workers, it makes sense). The good news is that we noticed no overhead (MSGPACK - REDIS) > > > > Sorry if it is still a custom code, we plan to discuss all this stuff in > > TUB 2016 and hopefully get part of them in tryton out of the box. > Will be great if we can discuss on the ML/BT/discuss server so other > peopole not attending at TUB can also join the discussion ;-) > Sure ! I just need to work on a presentation and a small demonstration to get things ready to show (I am making the effort for TUB) Discussions will continue later with all the community > > -- > Sergi Almacellas Abellana > www.koolpi.com > Twitter: @pokoli_srk > -- You received this message because you are subscribed to the Google Groups "tryton-dev" group. To view this discussion on the web visit https://groups.google.com/d/msgid/tryton-dev/9ceb1530-43a5-4611-b5d2-6ee243c902e3%40googlegroups.com.
