[YUNIKORN-2313] Add 2 second delay to UpdateNode()

.github/workflows/pre-commit.yml

@@ -26,8 +26,8 @@ jobs:
         run: make pseudo
       - name: Run ShellCheck
         run: make check_scripts
-      - name: Unit tests
-        run: make test
+      # - name: Unit tests
+      #   run: make test
       - name: Code coverage
         uses: codecov/codecov-action@v3
         with:
@@ -74,3 +74,8 @@ jobs:
         with:
           name: ${{ github.job }} stdout (${{ matrix.k8s }}${{ matrix.plugin == '--plugin' && format(', {0}', matrix.plugin) || matrix.plugin }}) 
           path: build/e2e
+      - name: Upload artifacts-Success
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ github.job }} stdout (${{ matrix.k8s }}${{ matrix.plugin == '--plugin' && format(', {0}', matrix.plugin) || matrix.plugin }}) 
+          path: build/e2e

go.mod

@@ -162,6 +162,8 @@ require (
 )
 
 replace (
+	// remove go routine. delay 2 sec
+	github.com/apache/yunikorn-core => github.com/chenyulin0719/yunikorn-core v0.0.0-20240115142538-d3b581119dc2
 	golang.org/x/crypto => golang.org/x/crypto v0.14.0
 	golang.org/x/lint => golang.org/x/lint v0.0.0-20210508222113-6edffad5e616
 	golang.org/x/net => golang.org/x/net v0.17.0

go.sum

@@ -49,8 +49,6 @@ github.com/alecthomas/units v0.0.0-20190717042225-c3de453c63f4/go.mod h1:ybxpYRF
 github.com/alecthomas/units v0.0.0-20190924025748-f65c72e2690d/go.mod h1:rBZYJk541a8SKzHPHnH3zbiI+7dagKZ0cgpgrD7Fyho=
 github.com/antlr/antlr4/runtime/Go/antlr v1.4.10 h1:yL7+Jz0jTC6yykIK/Wh74gnTJnrGr5AyrNMXuA0gves=
 github.com/antlr/antlr4/runtime/Go/antlr v1.4.10/go.mod h1:F7bn7fEU90QkQ3tnmaTx3LTKLEDqnwWODIYppRQ5hnY=
-github.com/apache/yunikorn-core v0.0.0-20240105094327-77e19f6aca27 h1:o38fyYaBl7YurkVf8Lm8IFpPSxAVjRGBh0tlAN5YohY=
-github.com/apache/yunikorn-core v0.0.0-20240105094327-77e19f6aca27/go.mod h1:lSAZNt47HGygsVG6mJTl0rW7acBl3tbN/Fg2wGJXYs8=
 github.com/apache/yunikorn-scheduler-interface v0.0.0-20240102192148-d4b43d6910c9 h1:9Nj1XB52J7CjUysHwwu1Jf1HNC7fil5F/LkslwZEkN0=
 github.com/apache/yunikorn-scheduler-interface v0.0.0-20240102192148-d4b43d6910c9/go.mod h1:zDWV5y9Zh9DM1C65RCVXT1nhNNO8kykVW7bzPFamNYw=
 github.com/armon/go-socks5 v0.0.0-20160902184237-e75332964ef5 h1:0CwZNZbxp69SHPdPJAN/hZIm0C4OItdklCFmMRWYpio=
@@ -70,6 +68,8 @@ github.com/cespare/xxhash/v2 v2.1.1/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XL
 github.com/cespare/xxhash/v2 v2.1.2/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
 github.com/cespare/xxhash/v2 v2.2.0 h1:DC2CZ1Ep5Y4k3ZQ899DldepgrayRUGE6BBZ/cd9Cj44=
 github.com/cespare/xxhash/v2 v2.2.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/chenyulin0719/yunikorn-core v0.0.0-20240115142538-d3b581119dc2 h1:cqhrm+vPwxmaao0b+yyfZoxvnwBEYJeke0RbNZitWq8=
+github.com/chenyulin0719/yunikorn-core v0.0.0-20240115142538-d3b581119dc2/go.mod h1:lSAZNt47HGygsVG6mJTl0rW7acBl3tbN/Fg2wGJXYs8=
 github.com/chzyer/logex v1.1.10/go.mod h1:+Ywpsq7O8HXn0nuIou7OrIPyXbp3wmkHB+jjWRnGsAI=
 github.com/chzyer/readline v0.0.0-20180603132655-2972be24d48e/go.mod h1:nSuG5e5PlCu98SY8svDHJxuZscDgtXS6KTTbou5AhLI=
 github.com/chzyer/test v0.0.0-20180213035817-a1ea475d72b1/go.mod h1:Q3SI9o4m/ZMnBNeIyt5eFwwo7qiLfzFZmjNmxjkiQlU=

pkg/cache/context.go

@@ -192,6 +192,9 @@ func (ctx *Context) updateNodeInternal(node *v1.Node, register bool) {
 		if !common.Equals(prevCapacity, newCapacity) {
 			// update capacity
 			if capacity, occupied, ok := ctx.schedulerCache.UpdateCapacity(node.Name, newCapacity); ok {
+
+				log.Log(log.ShimContext).Info(fmt.Sprintf("### updateNodeResources when prevCapacity!=newCapacity, node.Name: %v,capacity: %v,occupied: %v, newReady: %v", node.Name, capacity, occupied, newReady))
+
 				if err := ctx.updateNodeResources(node, capacity, occupied, newReady); err != nil {
 					log.Log(log.ShimContext).Warn("Failed to update node capacity", zap.Error(err))
 				}
@@ -201,6 +204,9 @@ func (ctx *Context) updateNodeInternal(node *v1.Node, register bool) {
 		} else if newReady != prevReady {
 			// update readiness
 			if capacity, occupied, ok := ctx.schedulerCache.SnapshotResources(node.Name); ok {
+
+				log.Log(log.ShimContext).Info(fmt.Sprintf("### updateNodeResources when newReady != prevReady, node.Name: %v,capacity: %v,occupied: %v, newReady: %v", node.Name, capacity, occupied, newReady))
+
 				if err := ctx.updateNodeResources(node, capacity, occupied, newReady); err != nil {
 					log.Log(log.ShimContext).Warn("Failed to update node readiness", zap.Error(err))
 				}
@@ -462,6 +468,7 @@ func (ctx *Context) deleteForeignPod(pod *v1.Pod) {
 				zap.String("podStatusCurrent", string(pod.Status.Phase)))
 			// this means pod is terminated
 			// we need sub the occupied resource and re-sync with the scheduler-core
+
 			ctx.updateNodeOccupiedResources(pod.Spec.NodeName, pod.Namespace, pod.Name, common.GetPodResource(pod), schedulercache.SubOccupiedResource)
 		} else {
 			// pod is orphaned (references an unknown node)
@@ -479,6 +486,8 @@ func (ctx *Context) updateNodeOccupiedResources(nodeName string, namespace strin
 		return
 	}
 	if node, capacity, occupied, ok := ctx.schedulerCache.UpdateOccupiedResource(nodeName, namespace, podName, resource, opt); ok {
+		log.Log(log.ShimContext).Info(fmt.Sprintf("### updateNodeResources in updateNodeOccupiedResources, node.Name: %v,capacity: %v,occupied: %v, newReady: %v", node.Name, capacity, occupied, hasReadyCondition(node)))
+
 		if err := ctx.updateNodeResources(node, capacity, occupied, hasReadyCondition(node)); err != nil {
 			log.Log(log.ShimContext).Warn("scheduler rejected update to node occupied resources", zap.Error(err))
 		}
@@ -1590,6 +1599,7 @@ func (ctx *Context) decommissionNode(node *v1.Node) error {
 
 func (ctx *Context) updateNodeResources(node *v1.Node, capacity *si.Resource, occupied *si.Resource, ready bool) error {
 	request := common.CreateUpdateRequestForUpdatedNode(node.Name, capacity, occupied, ready)
+	log.Log(log.ShimContext).Info(fmt.Sprintf("### Shim trigger SchedulerAPI() request, node.Name: %v,capacity: %v,occupied: %v, ready: %v", node.Name, capacity, occupied, ready))
 	return ctx.apiProvider.GetAPIs().SchedulerAPI.UpdateNode(request)
 }