ci: fix memory leak reporting caused by ovn-controller crashes (#3873)

Signed-off-by: zhangzujian <zhangzujian.7@gmail.com>
kubeovn · Mar 26, 2024 · 9b183d2 · 9b183d2
1 parent 720e5f6
commit 9b183d2
Showing 1 changed file with 12 additions and 8 deletions.
diff --git a/.github/workflows/build-x86-image.yaml b/.github/workflows/build-x86-image.yaml
@@ -522,10 +522,11 @@ jobs:
 
       - name: Check valgrind result
         run: |
-          kubectl -n kube-system rollout restart deploy ovn-central
           kubectl -n kube-system rollout restart ds ovs-ovn
-          kubectl -n kube-system rollout status deploy ovn-central
           kubectl -n kube-system rollout status ds ovs-ovn
+          sleep 10
+          kubectl -n kube-system rollout restart deploy ovn-central
+          kubectl -n kube-system rollout status deploy ovn-central
           while true; do
             if [ $(kubectl -n kube-system get pod -l app=ovs -o name | wc -l) -eq $(kubectl get node -o name | wc -l) ]; then
               break
@@ -678,10 +679,11 @@ jobs:
 
       - name: Check valgrind result
         run: |
-          kubectl -n kube-system rollout restart deploy ovn-central
           kubectl -n kube-system rollout restart ds ovs-ovn
-          kubectl -n kube-system rollout status deploy ovn-central
           kubectl -n kube-system rollout status ds ovs-ovn
+          sleep 10
+          kubectl -n kube-system rollout restart deploy ovn-central
+          kubectl -n kube-system rollout status deploy ovn-central
           while true; do
             if [ $(kubectl -n kube-system get pod -l app=ovs -o name | wc -l) -eq $(kubectl get node -o name | wc -l) ]; then
               break
@@ -955,10 +957,11 @@ jobs:
 
       - name: Check valgrind result
         run: |
-          kubectl -n kube-system rollout restart deploy ovn-central
           kubectl -n kube-system rollout restart ds ovs-ovn
-          kubectl -n kube-system rollout status deploy ovn-central
           kubectl -n kube-system rollout status ds ovs-ovn
+          sleep 10
+          kubectl -n kube-system rollout restart deploy ovn-central
+          kubectl -n kube-system rollout status deploy ovn-central
           while true; do
             if [ $(kubectl -n kube-system get pod -l app=ovs -o name | wc -l) -eq $(kubectl get node -o name | wc -l) ]; then
               break
@@ -1109,10 +1112,11 @@ jobs:
 
       - name: Check valgrind result
         run: |
-          kubectl -n kube-system rollout restart deploy ovn-central
           kubectl -n kube-system rollout restart ds ovs-ovn
-          kubectl -n kube-system rollout status deploy ovn-central
           kubectl -n kube-system rollout status ds ovs-ovn
+          sleep 10
+          kubectl -n kube-system rollout restart deploy ovn-central
+          kubectl -n kube-system rollout status deploy ovn-central
           while true; do
             if [ $(kubectl -n kube-system get pod -l app=ovs -o name | wc -l) -eq $(kubectl get node -o name | wc -l) ]; then
               break