Merge pull request #33787 from madhusudancs/fed-sleep-cleanup

Automatic merge from submit-queue

[Federation] Sleep for 2m (arbitrarily chosen) to let the underlying clusters to clean up all the resources they are holding.

cc @kubernetes/sig-cluster-federation @colhom @nikhiljindal

Ref: Issue #33388, Issue #28615, Issue #31624.
This commit is contained in:
Kubernetes Submit Queue 2016-09-29 15:51:03 -07:00 committed by GitHub
commit 5f4c8c241e

View File

@ -32,10 +32,20 @@ prepare-e2e
if [[ "${FEDERATION:-}" == "true" ]];then if [[ "${FEDERATION:-}" == "true" ]];then
source "${KUBE_ROOT}/federation/cluster/common.sh" source "${KUBE_ROOT}/federation/cluster/common.sh"
for zone in ${E2E_ZONES};do for zone in ${E2E_ZONES};do
# bring up e2e cluster # bring down an e2e cluster
( (
set-federation-zone-vars "$zone" set-federation-zone-vars "$zone"
cleanup-federation-api-objects || echo "Couldn't cleanup federation api objects" cleanup-federation-api-objects || echo "Couldn't cleanup federation api objects"
# TODO(madhusudancs): This is an arbitrary amount of sleep to give Kubernetes
# clusters enough time to delete the underlying cloud provider resources
# corresponding to the Kubernetes resources we deleted as part of the test
# teardowns. It is shameful that we are doing this, but this is just a bandage
# to stop the bleeding. Please don't use this pattern anywhere. Remove this
# when proper cloud provider cleanups are implemented in the individual test
# `AfterEach` blocks.
sleep 2m
test-teardown test-teardown
) )
done done