Add section describing how to run for development

[LucasRoesler]

What

• Add a Development section that describes how to run and test the certifier locally using faas-netes and KinD. This should make it easier for people to contibute because it will standardize the workflow for verifying a contribution.

Resolves #46

Signed-off-by: Lucas Roesler roesler.lucas@gmail.com

[alexellis]

space-mini:certifier alex$ make test-kuberneteserror: unable to forward port because pod is not running. Current status=Pending

Can you add one command into that block like?

kubectl rollout status -n openfaas deploy/gateway

[alexellis]

Interestingly enough I am getting errors like this from KinD recently:

kube-system          115s        Warning   FailedScheduling          pod/coredns-6955765f44-s2r8t                   0/1 nodes are available: 1 node(s) had taints that the pod didn't tolerate.

[alexellis]

Have the options changed for Swarm, or are those still correct? Should we include swarm leave / init so that it's more consistent with the new K8s instructions?

[alexellis]

The hey step is generating a lot of spam output like:

Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080
Handling connection for 8080

I have so much scroll-back that I can't find the earlier test results.

If you use keep alive or fewer connections there should be fewer statements. We could also pipe port-forward to dev null?

[alexellis]

--- PASS: Test_ScaleUpAndDownFromThroughPut (164.09s)

Is there a way we can use sed or helm chart flags to speed up the mechanism to show that it works without adding 2.7minutes of CI time?

[LucasRoesler]

--- PASS: Test_ScaleUpAndDownFromThroughPut (164.09s)

Is there a way we can use sed or helm chart flags to speed up the mechanism to show that it works without adding 2.7minutes of CI time?

This PR https://github.com/openfaas/certifier/pull/34 specifically attempts to address this without needing any extra sed magic

[LucasRoesler]

Have the options changed for Swarm, or are those still correct? Should we include swarm leave / init so that it's more consistent with the new K8s instructions?

The new instructions are not for "how to run it with kubernetes" they are specific to the certifier development flow. The sections for running for kubernetes and swarm have not been touched and as far as I know there is not special reason to change it. I don't think this readme is the place to add the instructions for how to deploy and run faas-swarm

[alexellis]

Do they still work as expected with Swarm after the recent changes?