G-Harmon
commented
6 years ago mostly looks good. just a few minor comments.
Reviewed 3 of 3 files at r1. Review status: all files reviewed at latest revision, all discussions resolved.
app/kubemci/cmd/delete.go, line 130 at r1 (raw file):
return ingErr } fmt.Printf("%s. Continuing with force delete", ingErr)
Is it really a good idea to continue if UnmarshallAndApplyDefault fails? It just checks the namespaces and the annotations.
app/kubemci/cmd/delete.go, line 144 at r1 (raw file):
return clientsErr } fmt.Printf("%s. Continuing with force delete", clientsErr)
I think you need a newline at the end of Printf()s
app/kubemci/pkg/gcp/loadbalancer/loadbalancersyncer.go, line 231 at r1 (raw file):
// DeleteLoadBalancer deletes the GCP resources associated with the L7 GCP load balancer for the given ingress. // TODO(nikhiljindal): Do not require the ingress yaml from users. Just the name should be enough. We can fetch ingress YAML from one of the clusters. func (l *LoadBalancerSyncer) DeleteLoadBalancer(ing *v1beta1.Ingress, forceDelete bool) error {
nit: document new parameters.
app/kubemci/pkg/gcp/loadbalancer/loadbalancersyncer_test.go, line 405 at r1 (raw file):
t.Fatalf("unexpected error in test setup: %s", err) } if err := lbc.CreateLoadBalancer(ing, true /*forceUpdate*/, true /*validate*/, clusters); err != nil {
nit: set force=false while creating?
Comments from Reviewable
nikhiljindal
Fixes https://github.com/GoogleCloudPlatform/k8s-multicluster-ingress/issues/164
Adding a
--forceflag to kubemci delete to continue deletion even in case of errors.Will add an e2e test as part of remove-clusters test (delete the mci after it has been removed from all clusters)
cc @G-Harmon
This change is