You might want to try basically disabling redis by setting the cache expiration to 5 seconds or some ultra-low number. We didn’t see any negative impacts from doing this, but ymmv. Note that Mongo will cache things anyway, so redis may be redundant except under load/large installations.
--- Manny Veloso / technical person o: +360 859 1780 Follow SmartRG <http://www.smartrg.com/> news and events on Facebook <https://www.facebook.com/SmartRG?fref=ts> | Twitter <https://twitter.com/SRGcrew> | LinkedIn <https://www.linkedin.com/company/2568791?trk=vsrp_companies_res_name&trkInfo=VSRPsearchId%3A80173221376409258784%2CVSRPtargetId%3A2568791%2CVSRPcmpt%3Aprimary> | YouTube <https://www.youtube.com/channel/UCxs7PruCYyHnuAVHO_oQAmQ> On 4/12/16, 6:19 PM, "Users on behalf of Zaid Abdulla" <[email protected] on behalf of [email protected]> wrote: >On Tue, 2016-04-12 at 08:57 +0100, James Harrison wrote: >> Can GenieACS run with two masters behind a loadbal like HAproxy >> without shared Redis state, put another way? > >Should be doable with a small change in the lib/nbi.coffee to ensure >any cache invalidation resulting from API calls takes place on all >redis nodes. You can search for "db.redisClient" in that file to find >the lines that need to be changed. > >You'll also want to configure your proxy server such that requests from >a given client always go to the same node in order to reduce cache >miss. > >Zaid >_______________________________________________ >Users mailing list >[email protected] >http://lists.genieacs.com/mailman/listinfo/users _______________________________________________ Users mailing list [email protected] http://lists.genieacs.com/mailman/listinfo/users
