On Wed, 15 Sept 2021 at 19:43, Aom Jeff Root <aomjeff.r...@gmail.com> wrote:
> Hi Mr Maxim > >> *I hope my description makes sense :)* >> >> Of course after several tests performed and conclusion, I had the same > result. > > Thanks Mr Sebatian. > >> *Il y a eu une autre discussion récente (plus technique) sur les >> différents modes/questions de regroupement sur cette liste de diffusion qui >> pourrait valoir la peine d'être lue :* >> *https://markmail.org/message/w55hh7ri2tmtlnvt* >> <https://markmail.org/message/w55hh7ri2tmtlnvt> >> >> *Surtout aussi autour de l'aspect des tests de performance et de la >> mesure des limites de capacité. Ainsi que comment optimiser l'utilisation >> du matériel/réseau lorsque la demande augmente.* >> > > > with the cluster mode, each node has an url/ip. Can we configure > hazelcast.xml or elsewhere something (unique url) users will access and > they will not know if they are redirected or not. Like when one node is > down, they will continue work but on nodes available (up) ? To make > availability service. > This *might be* possible, but will complicate the configuration And will require to change source code and recompile > Kind regards, > Jeffrey. > > Le mer. 8 sept. 2021 à 21:57, seba.wag...@gmail.com <seba.wag...@gmail.com> > a écrit : > >> There was another recent (more technical) discussion about different >> clustering modes/questions on this mailing list that might be worth reading: >> https://markmail.org/message/w55hh7ri2tmtlnvt >> >> Especially also around the performance testing aspect >> and measuring capacity limits. As well as how to optimise the >> hardware/network usage when demand scales up. >> >> Thanks >> Sebastian >> >> Sebastian Wagner >> Director Arrakeen Solutions, OM-Hosting.com >> http://arrakeen-solutions.co.nz/ >> https://om-hosting.com - Cloud & Server Hosting for HTML5 >> Video-Conferencing OpenMeetings >> >> <https://www.youracclaim.com/badges/da4e8828-743d-4968-af6f-49033f10d60a/public_url> >> <https://www.youracclaim.com/badges/b7e709c6-aa87-4b02-9faf-099038475e36/public_url> >> >> >> On Thu, 9 Sept 2021 at 03:27, Maxim Solodovnik <solomax...@gmail.com> >> wrote: >> >>> The current balancer is hardcoded and have no options for tuning (maybe >>> yet, ideas are welcome) >>> >>> The clustering is room-based >>> So if you have 2 nodes >>> >>> First user accessing the roomA will use node1, >>> all other users entering roomA will be redirected to node1 >>> >>> If userB will enter roomB he/she will be redirected to node2 >>> As well as all other users entering roomB >>> >>> if userC will enter roomC - the clustering logic will check >>> 1) if free nodes exists >>> 2) IF NOT: roomA and roomB capacity will be checked >>> Node with smaller total capacity will be used >>> >>> i.e. if roomA is personal room for max. 15 users and roomB is public >>> room with max. 150 users >>> roomC will be opened at node1 >>> >>> I hope my description makes sense :) >>> >>> On Fri, 3 Sept 2021 at 09:46, Aom Jeff Root <aomjeff.r...@gmail.com> >>> wrote: >>> >>>> Hello, >>>> I have some questions about cluster of om >>>> >>>> In cluster mode, how is the redirection of users to the room >>>> appropriate for the server? >>>> What are the criteria? >>>> Are there any configurations to be made at the cluster level to know >>>> when it should redirect users ? >>>> We have which advantages with the cluster with 2 or more om ? >>>> >>>> thanks in advance. >>>> >>>> Kind regards, >>>> Jeffrey >>>> >>>> Le jeu. 2 sept. 2021 à 06:29, Aom Jeff Root <aomjeff.r...@gmail.com> a >>>> écrit : >>>> >>>>> In the scenario where the second instance is not turned off manually. >>>>> I.e. they are in a meeting and suddenly he had a crash users connected to >>>>> this instance will not have access to the application (i.e. the server >>>>> will >>>>> be offline) >>>>> >>>>> Le jeu. 2 sept. 2021 à 03:17, Maxim Solodovnik <solomax...@gmail.com> >>>>> a écrit : >>>>> >>>>>> Hello Jeffrey, >>>>>> >>>>>> since taking cluster node offline is manual process I would expect it >>>>>> will happen while there will be no load >>>>>> I can imagine no mechanism to more users to other node >>>>>> >>>>>> might require some "Cluster admin" or something like this .... >>>>>> >>>>>> On Thu, 2 Sept 2021 at 01:07, Aom Jeff Root <aomjeff.r...@gmail.com> >>>>>> wrote: >>>>>> >>>>>>> Hi dear all, >>>>>>> Need help - clarifications on setting up cluster - OM >>>>>>> >>>>>>> When two OM instances are in cluster mode, is it normal that if one >>>>>>> is turned off, the meeting (access) is no longer available for the users >>>>>>> connected to this instance? >>>>>>> Isn't there a way to switch them automatically to the instance that >>>>>>> is active? >>>>>>> >>>>>>> Best regards, >>>>>>> Jeffrey >>>>>>> >>>>>>> Le mar. 31 août 2021 à 06:12, Aom Jeff Root <aomjeff.r...@gmail.com> >>>>>>> a écrit : >>>>>>> >>>>>>>> Hi dear all, >>>>>>>> >>>>>>>> When two OM instances are in cluster mode, is it normal that if one >>>>>>>> is turned off, the meeting (access) is no longer available for the >>>>>>>> users >>>>>>>> connected to this instance? >>>>>>>> Isn't there a way to switch them to the instance that is >>>>>>>> automatically active? >>>>>>>> >>>>>>>> Best regards, >>>>>>>> Jeffrey >>>>>>>> >>>>>>> >>>>>> >>>>>> -- >>>>>> Best regards, >>>>>> Maxim >>>>>> >>>>> >>> >>> -- >>> Best regards, >>> Maxim >>> >> -- Best regards, Maxim