From 7a930fa40204afb273e3823a6a6659d402f177a2 Mon Sep 17 00:00:00 2001 From: Marco Hofstetter Date: Tue, 14 May 2024 12:32:58 +0200 Subject: [PATCH] ci: increase wait duration after upgrade/downgrade in E2E upgrade test E2E Upgrade tests randomly fail if health endpoints aren't available after an upgrade within the given default timeout of 5min. ``` Errors: cilium cilium-c4qkb controller cilium-health-ep is failing since 2m7s (15x): Get "http://10.244.2.186:4240/hello": dial tcp 10.244.2.186:4240: connect: no route to host ``` It looks like the health server unix listener isn't available at that time. Therefore, this commit increases the timeout from 5 to 10 minutes. Suggested-by: Marco Iorio Signed-off-by: Marco Hofstetter --- .github/workflows/tests-e2e-upgrade.yaml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/tests-e2e-upgrade.yaml b/.github/workflows/tests-e2e-upgrade.yaml index 1fb6acd616855..1b02b4ac5e4d6 100644 --- a/.github/workflows/tests-e2e-upgrade.yaml +++ b/.github/workflows/tests-e2e-upgrade.yaml @@ -432,7 +432,7 @@ jobs: ./cilium-cli upgrade \ ${{ steps.cilium-newest-config.outputs.config }} - ./cilium-cli status --wait + ./cilium-cli status --wait --wait-duration=10m kubectl get pods --all-namespaces -o wide kubectl -n kube-system exec daemonset/cilium -- cilium status @@ -463,7 +463,7 @@ jobs: ./cilium-cli upgrade \ ${{ steps.cilium-stable-config.outputs.config }} - ./cilium-cli status --wait + ./cilium-cli status --wait --wait-duration=10m kubectl get pods --all-namespaces -o wide kubectl -n kube-system exec daemonset/cilium -- cilium status