K8s Service redeployment

[manusa]

Fixes #19701

Deploy step checks if resource needs to be deleted before creation. In case deletion is needed (KNATIVE or Service, others should be opted-in in the new shouldDeleteExisting method), client deletes any existing resource matching Group+Version+Kind+Namespace+Name and waits for its deletion (10 seconds ---> Maybe should add a constant somewhere?)

I'd like to add some tests for this, but I'm not sure where.

/cc @geoand

[quarkus-bot]

Thanks for your pull request!

The title of your pull request does not follow our editorial rules. Could you have a look?

• title should preferably start with an uppercase character (if it makes sense!)
• title should not start with chore/docs/feat/fix/refactor but be a proper sentence

This message is automatically generated by a bot.

[iocanel]

LGTM

[geoand]

I'd like to add some tests for this, but I'm not sure where.

For the time being, we don't have a cluster are part of our upstream CI.

@Sgitario it would be great if you could test this at some point

[Sgitario]

I'd like to add some tests for this, but I'm not sure where.

For the time being, we don't have a cluster are part of our upstream CI.

@Sgitario it would be great if you could test this at some point

I'd like to add some tests for this, but I'm not sure where.

For the time being, we don't have a cluster are part of our upstream CI.

@Sgitario it would be great if you could test this at some point

This is a good one! I will add an automated test into our repo https://github.com/quarkus-qe/quarkus-test-suite to reproduce this issue using https://github.com/quarkus-qe/quarkus-test-framework#kubernetes

[geoand]

Thanks!

[gsmet]

@Sgitario any chance you can validate this one before next Monday? It would be nice to have it in 2.2.2.Final.

[Sgitario]

@Sgitario any chance you can validate this one before next Monday? It would be nice to have it in 2.2.2.Final.

@geoand @gsmet I could not verify the issue at first place: #19701 (comment)

[geoand]

So you were able to redeploy the application without problems?

[Sgitario]

So you were able to redeploy the application without problems?

Yes, without the changes from this PR.

[geoand]

Weird... I think I was able to reproduce the problem in Minikube

[manusa]

So you were able to redeploy the application without problems?

I did reproduce the issue too (and obviously I verified the fix) in Minikube. It's a common problem that you can easily reproduce even using kubectl with the generated YAMLs (since there is an already allocated node port/IP, the service cannot start).

What I'm not sure if the issue is reproducible in the 1.x LTS version, since in some versions of the client a specific Service handler might have taken care of this.

[Sgitario]

Weird... I think I was able to reproduce the problem in Minikube

We use kind and OCP, but not minikube.
As @manusa said, if he could reproduce it in minikube, then I guess it should be fine.

[geoand]

Let's merge this as it clearly fixes a problem and doesn't seem to cause any new ones.

Thanks folks