[OOTB-hive] Thoughts on Honeycomb that need your input

[Andreas Steffan]

Scaling has various aspects. Alfresco will surely not throw application
level clustering code at us. From what I observe, I'd say they are
already slicing and dicing code and artifacts in various ways in order
to get more flexible at (automatic) plumbing. They need to get
installation to the next level, and the executable installer won't cut
it. I am fairly sure the community will see more benefit from those efforts.

However, I don't think this really matters today if we are aiming at
quickly getting something usable out of the door.

Same goes for jenkins and/or artifactory. Really awesome, especially at
scale but imho not critical in order to get something out of the door.

A basic webdav enabled webserver could do for a start. Maybe we could
even ask Alfresco whether we could use artifacts.alfresco.com. ;)
Seriously, I don't see why this should be a big deal.

Again, personally I'd like to start with things being as simple as possible.

Running Honeycomb should imho be as simple as:

docker run ootb/honeycomb:1.0

Running plain CE should be:

docker run alfresco/alfresco-community:5.1.g

PS: I guess I will be putting my thoughts on Alfresco containerization
together and push another blog post out.

cheers
Andreas
 

On 07/13/2016 12:47 PM, martin at bettercode.com wrote:
> Hehe, well I figured I wanted to get the whole story down in one place
> :-)  Unfortunately halfway through I obviously thought I was in the
> INFRA channel! :-/
>
> As far as I see it, Honeycomb *is just* a set of extensions, rolled into
> a vanilla Alfresco CE, at least as it was originally stated back when we
> started, and recently re-stated by Jeff. The puppet thing started out as
> a build around which to test said extensions and just kind of
> snowballed. That said, we're in a better position now to have something
> that we can say *is not* Honeycomb (even though we temporarily
> appropriated the name) than if there was nothing there at all right now.
>  At least we have something to talk about :-)
>
> Let's be serious though, Alfresco are never going to help us in anything
> that looks like it might help CE scale, it would be sailing too close to
> their enterprise offering, so I doubt we are going to see anything
> helpful there. Like I said, marsbard/docker-alfresco is merely our
> (Daren's and I) experimentation in that direction.
>
> Developing with Docker though is super nice compared to puppet. Once you
> get a layer locked in there is no more need to "start from scratch", so
> the development lifecycle is really fast. It's also really nice that the
> end result is a text file, the Dockerfile, and also that you can (but I
> didn't) publish the completed image to the (or a) Docker Hub.
>
> We need the build process (i.e. Jenkins) to build not only the
> extensions but also alfresco.war, share.war et al and then store them in
> our own artifact server. From there we are free to have whatever
> installer we want, be it a Dockerfile, puppet manifests, shell scripts
> or whatever. Perhaps for example Loftux would prefer to get their war
> artifacts from us rather than from potential zombie Patient Zero. I'm
> definitely not very keen on using the alfresco-provided installers. So
> basically it's all still blocked by the infrastructure. (But you were
> right, back then 2 years ago when you said that Docker could have been a
> better approach than puppet. If a Dockerfile had appeared back then we
> could be having a very different conversation now.)
>
> Cheers
> Martin
>
> PS Thanks for your kind comment about my health


-- 
Andreas Steffan

Achter Billing 14
22399 Hamburg
Germany

skype: deas0815
M: +49 160 4694826
T: +49 40 23943542
F: +49 40 23943542

http://www.contentreich.de

Contentreich : Alfresco ECM, Clojure, Groovy und WordPress - aus Spaß und für Geld