Project/Namespace dependencies on the cluster being active

[josmo]

Project/Namespace dependencies on the cluster being active

[rawmind0]

This is not related with issue #12. rancher2_namespace and rancher2_project resources need that the cluster is active to be created.

Submitted PR #73, that adds wait_for_cluster argument on rancher2_namespace and rancher2_project resources to wait for cluster is active.

[jnicholls]

See my comment on PR #73. Beyond the rancher2_namespace and rancher2_project resources, there are other kubernetes-based resources which rely on the kube-apiserver being ready. It would be great to create a new type of resource named something like rancher2_cluster_wait_until_active that takes the cluster_id and an optional timeout, and will wait until the cluster's kube-apiserver becomes ready before proceeding. Then, kubernetes/helm/etc. resources can explicitly be set to depend upon that resource and will thus wait for it to return.

Thoughts?

[josmo]

Thanks @rawmind0!!

[josmo]

@jnicholls I like the proposal, I don't think it's a huge deal for namespaces and the likes in k8s resources since you could still use the rancher resources for that, however things like setting up volume classes that doesn't have deps on namespaces, starting in the system/default namesspaces, it would be really useful

[narfeta]

hi all, any update about this? it would be really useful to have this feature.

[drpebcak]

I think it makes sense to have optional waits on all the rancher2 resources, but then also have something separate that can tell you if the cluster itself is ready, like rancher2_cluster_wait_until_active

[rawmind0]

I think that @jnicholls proposal is indeed moreover to "external" rancher2 resources. For rancher2 resources, not sure it it's needed on all, but PR #73 extra argument should be enough, due to they are able to check the cluster state.

[rawmind0]

Merged PR that address this issue. Please, reopen issue if needed.

@jnicholls, could you please open a new issue if you are still interested on your rancher2_cluster_wait_until_active proposal??