This version (2017/05/27 13:44) is a draft.
Approvals: 0/1

[11:53:35] *** ChanServ sets mode: +o temporalfox

[13:44:42] <skullcrasher> hey :) I just started looking through the vertx docs. I have a question regarding “deployment” of a vertx application.

[13:44:55] <skullcrasher> Is there some mechanism that allows adding further verticles on runtime

[16:50:26] <Sticky> skullcrasher: you can call Vertx.deployVerticle at any time

[17:05:52] <skullcrasher> Sticky, ok. But does it just include a new “feature” into my running vertx instance or start a new one. That's something I didn't fully get out of the docs

[17:06:40] <Sticky> not sure what you mean by feature

[17:06:46] <skullcrasher> Also another question is, how inter cluster communication is handled. Has this to do with the buffer and tcp-server section?

[17:07:02] <skullcrasher> Sticky, with feature I mean just a new functionality/service

[17:07:26] <Sticky> deployVerticle starts a new verticle

[17:08:05] <Sticky> so you could use verticle and service interchangibly

[17:09:22] <skullcrasher> ok nice.

[17:10:58] <skullcrasher> Anything on the inter cluster communication? So how can services in 2 different clusters talk to each other?

[17:11:39] <Sticky> I am not sure if there is anything for bridging different clusters

[17:11:48] <Sticky> not looked into it

[17:11:51] <skullcrasher> When looking at akka it seems quite easy to talk to totally different services.

[17:12:16] <skullcrasher> Sticky, ah ok. So that could be the part where the tcp client/server stuff comes in

[17:13:21] <Sticky> so a single cluster can be on multiple machines

[17:15:05] <Sticky> if you want multiple clusters and bridging traffic between them afaik you would have to implement that bridge yourself

[17:15:56] <skullcrasher> Sticky, ok, that's what I was talking about. how can multiple clusters (possible other hardware) communicate with eath other (or their services)

[17:16:05] <Sticky> but the question would be what is your usecase that you need multiple clusters?

[17:17:05] <skullcrasher> I don't think it has to be, because I think that the service grouping should be enough.

[17:18:29] <skullcrasher> But others here mentioned this could be a way to separate services from each other

[17:19:52] <Sticky> ahh, if its some thing like a security concern of doing something like splitting out client facing services and internal, thats a valid usecase

[17:20:34] <skullcrasher> Sticky, yes this is included in such a use case

[17:21:00] <skullcrasher> so only the “backend-cluster” talks to the “api-cluster” etc.

[17:21:11] <skullcrasher> and the services are separated

[17:21:43] <skullcrasher> there arises the questions on how to let the backend talk to the api etc.

[17:22:02] <skullcrasher> handle all ourselves with tcp servers in the verticles etc didn't look that nice

[17:22:03] <Sticky> yeah, afaik that cluster bridge does not exist out the box, but again I have never looked at this stuff so dont know

[17:22:31] <skullcrasher> Sticky, no problem :) thx a lot for answering the other question

[17:22:35] <skullcrasher> that helped a lot :)

[17:23:44] <Sticky> but to a large extent its going to be heavily specific to you, as you are going to want that bridge to be a firewall type thing that would prevent unwanted/unverified/unauthed etc data from jumping the bridge

[17:24:34] <Sticky> shouldnt be hard to build though

[17:26:14] <skullcrasher> hmm yeah. This is one of the “critical” desicions to make comparing akka that has it out of the box

[17:26:23] <skullcrasher> currently we evaluate wether akka or vertx to use

[17:26:55] <skullcrasher> vertx seems overall to feel more intuitive to use and has almost the exact features as akka.

[17:27:25] <skullcrasher> The only 2 differences we had until now were the dynamic service loading on runtime, that vertx has but akka doesn't.

[17:27:50] <skullcrasher> But on the other side the inter cluster communication that akka has but vertx doesn't

[17:27:51] <skullcrasher> :D

[17:27:52] <Sticky> yeah

[17:28:09] <Sticky> I would hang around to see if someone more informed knows of anythign

[17:28:45] <skullcrasher> yep

[17:28:48] <Sticky> but some quick googling I cant seem to find anything either

[17:28:56] <skullcrasher> yep same here :)

[17:29:04] <skullcrasher> that's why I ask here

[17:55:49] <iazarny> Hi All. Can I create cluster wide high available timer ?