Michele, I like the idea to make the ActionLoop based runtimes to be runnable on Knative.
My thoughts on this: - I second Markus concern to implement the invocation API onto Knative instead of just using Knative service syntax. - I would have concerns to make it dependent on Gloo which is kind of a minority choice for Knative load balancing - In my opinion the goal should be to have some uniform behaviour for ActionLoop based runtimes and other ones like the adapted NodeJS runtimes demonstrated by Matt and Priti - As Knative Build seems be on a dead end I would propose to target Tekton as the build system (which developed as kind of successor out of Knative) Maybe it would be a good solution to tackle two things independently. 1) Design and implement a common protocol of building, running and calling OW runtimes on Knative 2) Implement the OW invocation API on top of Knative as an additional option for those who have the need to expose it. I would looking forward to work with you on the first work item. Regards, Martin > On 20. May 2019, at 08:55, Michele Sciabarra <mich...@sciabarra.com> wrote: > > >>> I have an idea for implementing a prototype of OpenWhisk on top of Knative. >>> My basic ideas are: do not use any proxy, forwarding or adapter: extend >>> the runtime to support the REST call and expose them as ingress. And use a >>> wrapper on top of `kubectl` to generate all the needed components. > >> Does this tie into the work that Matt was doing to the runtimes to make >> them runnable on Knative? Is this lined up with that at all? > Actually yes. He suggested I can investigate how to migrate ActionLoop to > port many other languages to Knative. > Also he recommended I add my idea and this is what I am doing. Current code > is, if I am not wrong, a Knative build of the nodejs runtime. > > There has been a number of attempts and proposal to move forward OpenWhisk. > My idea that to succeed we need something small but that just works. This is > my idea to be able to implement in the shorter time frame possible an actual > subset of OpenWhisk that works and it is truly built on top of Knative. So I > am putting the thing a bit further than Matt work. > > >>> My goal is to have a functional work-alike of OpenWhisk built on top of >>> Knative, using ActionLoop as a foundation. I will extend ActionLoop to >>> support the required REST calls of OpenWhisk. >> >>> I also want to create tool, I will call `wskn`. This tool will initially >>> just a python script, a wrapper on top of `kubectl` as it will generate >>> kubernetes descriptors. >> Why not build this into "wsk" itself? The Azure Functions CLI as an example >> supports multiple deployment types like this in one CLI. > > When it will works, yes, of course. But to start, what I really need is a > prototype that can generate kubernetes descripttors to feed to kubectl, so a > simplee, quick and ditry, separate tool (that I will keep together the > runtime) is all I need for now. > >>> >>> It will support initially just the the action creation and invocation, and >>> only synchronous (blocking) behaviour, as all the request will go straight >>> to the runtimes. Hopefully also a subset of `package` and `activation`. >>> Again triggers, rules, asynchronous for later. >>> >>> The idea is that you will be able to create actions and web actions that >>> can run existing OpenWhisk actions, at least those with blocking behaviour >>> that run with ActionLoop (Go, Java, Python, PHP, Swift, Rust, Ruby, >>> Crystal...) >>> >>> Implementation. >>> ============== >>> >>> This is how I plan to implement it. >>> >>> At this stage I want to use just Knative Serving and Knative Build, using >>> Gloo for the ingress part. I also plan to install a local Docker registry >>> Kubernetes registry, so we do not have to use DockerHub for everything. All >>> of this can be done with existing command line tools in a few minutes in >>> any running Kubernetes deployment. >>> > >> Why specifying Gloo here? Do you need anything specific from Gloo itself? >> If not I'd propose to just keep it on a Knative Serving API surface level. > I want to build it on top of Knative serving, full stop. Currently, > installing Gloo is pretty easy and is more lightweight than Istio so I will > use it for my first implementation. > >>> >>> When I create an action, it will use Knative build that will work roughly >>> in this way: >>> >>> - create a configmap with the action code >>> - build the actin using ActionLoop precompilation feature that will return >>> a zip file including all the needed to run the action >>> - create a new docker image extending the runtime with the new zip, using >>> Kaanico >>> - push the image in the local registry >> This feels like a fairly heavyweight process, we should be able to come up >> with a way to circumvent zipping entirely. Maybe the runtime can detect >> that the unzipped content is already there and skip the unzip step? > > Actually this is my first idea of how to use Knative build. And is not > complicated: when I create the action, a run a build that includes Kanico. I > generate a Dockerfile on the fly. The docker file uses the action runtime > that already know how to compile a script. And then I save an image. I > already implemented un "autoinit" so just launching the image will give a > runtime ready to run that execute an action already compiled. > > >> I'm fairly hesitant on the usage of a ConfigMap for storing the action >> code. It's all stored in the in-cluster etcd instance and it has a limit of >> 1M. This is at most a stop-gap solution to provide a PoC I think. Any ideas >> on how to "productize" this? > > ConfigMap can be mounted as files, so it is an easy way to feed an action to > a build. It is just an easy way to feed the action code to the Build. > > My initial constraint is that I want just to generate Kubernetes descriptors > to feed to kubectl. > Of course in the long run I can add some "file upload" storage. > > If I could to this file upload when invoking a build it could ideal as I do > not have to store anything anywhere, just process the code and generate a > single layer to execute actions to be store in the registry. > I will investigate better this area, I understand your concern. > >> >>> At this point you can run the action. ActionLoop will be extended to >>> support invocations in the format >>> "/v1/namespaces/namespace/actions/package/action". >> Why bother reimplementing this exact path? To obtain API compatibility with >> OpenWhisk as it is today? > > I want to implement a subset of the OpenWhisk API on top of Knative serving. > Knative serving already does the scaling and routing, so what we need are the > "endpoints" to invoke actions. > > Since I do not want to add additional components, not at the first stage. > Just knative serve and build, the runtime and a controller script, the > runtime is the natural place where to "handle" the API invocations, since > Knative only generates the URL but not anything else. If I understood well, > Matt is adding a proxy. I do not want to add a proxy, just add to the runtime > the ability to respond to "API like" calls, at least those regarding action > invocation. > >>> It will do all the decoding required to invoke the action with the >>> expected paramenters (straight invocation thrhoug the actinloop protocol, >>> not proxies). >> Does this mean moving all of the Controller's "smartness" about incoming >> and outgoing HTTP requests (see the whole WebActions for example)? > > At least decoding web actions in the runtime, yes. Knative serving already > has routing and proxying. > So a true implementation on top of Knative requires IHMO this sacrifice. > Unless there is a way to keep the controller in a "Knative" compatible way. > Open to suggestions here. > >> Each action will then be exposed using an ingress with its specific >> invocation path. >> >> If the community agrees with this plan, I would create a repo >> `incubator-openwhisk-knative` to work on it. >> >> Thoughts?
