Updates to the controller logic to better handle failures in etc updates

config/crd/bases/mcad.ibm.com_appwrappers.yaml

@@ -1,5 +1,3 @@
-
----
 apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
@@ -778,6 +776,10 @@ spec:
                   QueueJob (by Informer)
                 format: date-time
                 type: string
+              controllerfirstdispatchtimestamp:
+                description: Microsecond level timestamp when controller first dispatches appwrapper
+                format: date-time
+                type: string
               failed:
                 description: The number of resources which reached phase Failed.
                 format: int32
@@ -790,8 +792,7 @@ spec:
                 description: Is Dispatched?
                 type: boolean
               local:
-                description: Indicate if message is a duplicate (for Informer to recognize
-                  duplicate messages)
+                description: Indicate if message is a duplicate (for Informer to recognize duplicate messages)
                 type: boolean
               message:
                 type: string
@@ -812,15 +813,13 @@ spec:
                 format: int32
                 type: integer
               queuejobstate:
-                description: State of QueueJob - Init, Queueing, HeadOfLine, Rejoining,
-                  ...
+                description: State of QueueJob - Init, Queueing, HeadOfLine, Rejoining ...
                 type: string
               running:
                 format: int32
                 type: integer
               sender:
-                description: Indicate sender of this message (extremely useful for
-                  debugging)
+                description: Indicate sender of this message (extremely useful for debugging)
                 type: string
               state:
                 description: State - Pending, Running, Failed, Deleted
@@ -834,9 +833,22 @@ spec:
                   (is this different from the MinAvailable from JobStatus)
                 format: int32
                 type: integer
+              numberOfRequeueings:
+                description: Field to keep track of how many times a requeuing event has been triggered
+                format: int32
+                type: integer
+                default: 0
+              requeueingTimeInSeconds:
+                description: Field to keep track of total number of seconds spent in requeueing 
+                format: int32
+                type: integer
+                default: 0
             type: object
         required:
         - spec
         type: object
     served: true
     storage: true
+    subresources:
+      status: {}
+

deployment/mcad-controller/crds/mcad.ibm.com_appwrappers.yaml

@@ -1,5 +1,3 @@
-
----
 apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
@@ -778,6 +776,10 @@ spec:
                   QueueJob (by Informer)
                 format: date-time
                 type: string
+              controllerfirstdispatchtimestamp:
+                description: Microsecond level timestamp when controller first dispatches appwrapper
+                format: date-time
+                type: string
               failed:
                 description: The number of resources which reached phase Failed.
                 format: int32
@@ -790,8 +792,7 @@ spec:
                 description: Is Dispatched?
                 type: boolean
               local:
-                description: Indicate if message is a duplicate (for Informer to recognize
-                  duplicate messages)
+                description: Indicate if message is a duplicate (for Informer to recognize duplicate messages)
                 type: boolean
               message:
                 type: string
@@ -812,15 +813,13 @@ spec:
                 format: int32
                 type: integer
               queuejobstate:
-                description: State of QueueJob - Init, Queueing, HeadOfLine, Rejoining,
-                  ...
+                description: State of QueueJob - Init, Queueing, HeadOfLine, Rejoining ...
                 type: string
               running:
                 format: int32
                 type: integer
               sender:
-                description: Indicate sender of this message (extremely useful for
-                  debugging)
+                description: Indicate sender of this message (extremely useful for debugging)
                 type: string
               state:
                 description: State - Pending, Running, Failed, Deleted
@@ -834,9 +833,22 @@ spec:
                   (is this different from the MinAvailable from JobStatus)
                 format: int32
                 type: integer
+              numberOfRequeueings:
+                description: Field to keep track of how many times a requeuing event has been triggered
+                format: int32
+                type: integer
+                default: 0
+              requeueingTimeInSeconds:
+                description: Field to keep track of total number of seconds spent in requeueing 
+                format: int32
+                type: integer
+                default: 0
             type: object
         required:
         - spec
         type: object
     served: true
     storage: true
+    subresources:
+      status: {}
+

go.mod

@@ -5,6 +5,7 @@ go 1.18
 require (
 	github.com/eapache/go-resiliency v1.3.0
 	github.com/emicklei/go-restful v2.16.0+incompatible
+	github.com/gogo/protobuf v1.3.1
 	github.com/golang/protobuf v1.4.3
 	github.com/hashicorp/go-multierror v1.1.1
 	github.com/kubernetes-sigs/custom-metrics-apiserver v0.0.0-20210311094424-0ca2b1909cdc
@@ -45,7 +46,6 @@ require (
 	github.com/go-openapi/jsonreference v0.19.5 // indirect
 	github.com/go-openapi/spec v0.20.0 // indirect
 	github.com/go-openapi/swag v0.19.12 // indirect
-	github.com/gogo/protobuf v1.3.1 // indirect
 	github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e // indirect
 	github.com/google/go-cmp v0.5.5 // indirect
 	github.com/google/gofuzz v1.1.0 // indirect

hack/run-e2e-kind.sh

@@ -34,6 +34,7 @@ export CLUSTER_CONTEXT="--name test"
 export IMAGE_ECHOSERVER="kicbase/echo-server:1.0"
 export IMAGE_UBUNTU_LATEST="ubuntu:latest"
 export IMAGE_UBI_LATEST="registry.access.redhat.com/ubi8/ubi:latest"
+export IMAGE_BUSY_BOX_LATEST="k8s.gcr.io/busybox:latest"
 export KIND_OPT=${KIND_OPT:=" --config ${ROOT_DIR}/hack/e2e-kind-config.yaml"}
 export KA_BIN=_output/bin
 export WAIT_TIME="20s"
@@ -207,27 +208,20 @@ function kind-up-cluster {
     exit 1
   fi
 
-  docker pull ${IMAGE_ECHOSERVER} 
-  if [ $? -ne 0 ]
-  then
-    echo "Failed to pull ${IMAGE_ECHOSERVER}"
-    exit 1
-  fi
-
-  docker pull ${IMAGE_UBUNTU_LATEST}
+  docker pull ${IMAGE_UBI_LATEST}
   if [ $? -ne 0 ]
   then
-    echo "Failed to pull ${IMAGE_UBUNTU_LATEST}"
+    echo "Failed to pull ${IMAGE_UBI_LATEST}"
     exit 1
   fi
-
-  docker pull ${IMAGE_UBI_LATEST}
+  
+  docker pull ${IMAGE_BUSY_BOX_LATEST}
   if [ $? -ne 0 ]
   then
-    echo "Failed to pull ${IMAGE_UBI_LATEST}"
+    echo "Failed to pull ${IMAGE_BUSY_BOX_LATEST}"
     exit 1
   fi
-
+ 
   if [[ "$MCAD_IMAGE_PULL_POLICY" = "Always" ]]
   then
     docker pull ${IMAGE_MCAD}
@@ -244,7 +238,7 @@ function kind-up-cluster {
   fi
   docker images
 
-  for image in ${IMAGE_ECHOSERVER} ${IMAGE_UBUNTU_LATEST} ${IMAGE_MCAD} ${IMAGE_UBI_LATEST}
+  for image in ${IMAGE_ECHOSERVER} ${IMAGE_UBUNTU_LATEST} ${IMAGE_MCAD} ${IMAGE_UBI_LATEST} ${IMAGE_BUSY_BOX_LATEST}
   do
     kind load docker-image ${image} ${CLUSTER_CONTEXT}
     if [ $? -ne 0 ]
@@ -330,8 +324,6 @@ function mcad-quota-management-down {
       echo "Failed to undeploy controller"
       exit 1
     fi
-    echo "Waiting for the test namespace to be cleaned up.."
-    sleep 60
 }
 
 function mcad-up {
@@ -402,4 +394,4 @@ setup-mcad-env
 kuttl-tests
 mcad-quota-management-down
 mcad-up
-go test ./test/e2e -v -timeout 120m -count=1
+go test ./test/e2e -v -timeout 130m -count=1

pkg/apis/controller/v1beta1/appwrapper.go

@@ -101,7 +101,7 @@ type AppWrapperService struct {
 }
 
 // AppWrapperResource is App Wrapper aggregation resource
-//todo: To be depricated
+// todo: To be depricated
 type AppWrapperResource struct {
 	metav1.TypeMeta   `json:",inline"`
 	metav1.ObjectMeta `json:"metadata"`
@@ -246,7 +246,7 @@ type AppWrapperStatus struct {
 	// Microsecond level timestamp when controller first sees QueueJob (by Informer)
 	ControllerFirstTimestamp metav1.MicroTime `json:"controllerfirsttimestamp,omitempty"`
 
-	// Microsecond level timestamp when controller first sets appwrapper in state Running
+	// Microsecond level timestamp when controller first dispatches appwrapper
 	ControllerFirstDispatchTimestamp metav1.MicroTime `json:"controllerfirstdispatchtimestamp,omitempty"`
 
 	// Tell Informer to ignore this update message (do not generate a controller event)
@@ -264,18 +264,25 @@ type AppWrapperStatus struct {
 	// Represents the latest available observations of pods under appwrapper
 	PendingPodConditions []PendingPodSpec `json:"pendingpodconditions"`
 
-	//Resources consumed
-
+	// Represents the number of cpu consumed by all pods belonging to an appwrapper.
 	TotalCPU float64 `json:"totalcpu,omitempty"`
 
+	// Represents the amount of memory consumed by all pods belonging to an appwrapper.
 	TotalMemory float64 `json:"totalmemory,omitempty"`
 
+	// Represents the total number of GPUs consumed by all pods belonging to an appwrapper.
 	TotalGPU int64 `json:"totalgpu,omitempty"`
+
+	// Field to keep track of total number of seconds spent in requeueing
+	RequeueingTimeInSeconds int `json:"requeueingTimeInSeconds,omitempty"`
+
+	// Field to keep track of how many times a requeuing event has been triggered
+	NumberOfRequeueings int `json:"numberOfRequeueings,omitempty"`
 }
 
 type AppWrapperState string
 
-//enqueued, active, deleting, succeeded, failed
+// enqueued, active, deleting, succeeded, failed
 const (
 	AppWrapperStateEnqueued              AppWrapperState = "Pending"
 	AppWrapperStateActive                AppWrapperState = "Running"

pkg/controller/clusterstate/cache/cache.go

@@ -334,7 +334,7 @@ func (sc *ClusterStateCache) updateState() error {
 }
 
 func (sc *ClusterStateCache) deleteJob(job *api.JobInfo) {
-	klog.V(4).Infof("[deleteJob] Attempting to delete Job <%v:%v/%v>", job.UID, job.Namespace, job.Name)
+	klog.V(10).Infof("[deleteJob] Attempting to delete Job <%v:%v/%v>", job.UID, job.Namespace, job.Name)
 
 	time.AfterFunc(5*time.Second, func() {
 		sc.deletedJobs.AddIfNotPresent(job)

pkg/controller/metrics/adapter/adapter.go

@@ -32,39 +32,40 @@ package adapter
 
 import (
 	"flag"
+	"net/http"
+	"os"
+
 	"github.com/project-codeflare/multi-cluster-app-dispatcher/cmd/kar-controllers/app/options"
 	openapinamer "k8s.io/apiserver/pkg/endpoints/openapi"
 	genericapiserver "k8s.io/apiserver/pkg/server"
-	"net/http"
-	"os"
 
 	"github.com/emicklei/go-restful"
 	"k8s.io/apimachinery/pkg/util/wait"
 	"k8s.io/client-go/rest"
 	"k8s.io/klog/v2"
 
-	adapterprov "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/metrics/adapter/provider"
 	"github.com/kubernetes-sigs/custom-metrics-apiserver/pkg/apiserver"
 	basecmd "github.com/kubernetes-sigs/custom-metrics-apiserver/pkg/cmd"
 	"github.com/kubernetes-sigs/custom-metrics-apiserver/pkg/provider"
 	generatedopenapi "github.com/kubernetes-sigs/custom-metrics-apiserver/test-adapter/generated/openapi"
+	adapterprov "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/metrics/adapter/provider"
 
 	clusterstatecache "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/clusterstate/cache"
 )
 
 // New returns a Cache implementation.
-func New(serverOptions *options.ServerOption, config *rest.Config, clusterStateCache clusterstatecache.Cache) *MetricsAdpater {
+func New(serverOptions *options.ServerOption, config *rest.Config, clusterStateCache clusterstatecache.Cache) *MetricsAdapter {
 	return newMetricsAdpater(serverOptions, config, clusterStateCache)
 }
 
-type MetricsAdpater struct {
+type MetricsAdapter struct {
 	basecmd.AdapterBase
 
 	// Message is printed on succesful startup
 	Message string
 }
 
-func (a *MetricsAdpater) makeProviderOrDie(clusterStateCache clusterstatecache.Cache) (provider.MetricsProvider, *restful.WebService) {
+func (a *MetricsAdapter) makeProviderOrDie(clusterStateCache clusterstatecache.Cache) (provider.MetricsProvider, *restful.WebService) {
 	klog.Infof("[makeProviderOrDie] Entered makeProviderOrDie()")
 	client, err := a.DynamicClient()
 	if err != nil {
@@ -79,7 +80,7 @@ func (a *MetricsAdpater) makeProviderOrDie(clusterStateCache clusterstatecache.C
 	return adapterprov.NewFakeProvider(client, mapper, clusterStateCache)
 }
 
-func covertServerOptionsToMetricsServerOptions(serverOptions *options.ServerOption) []string{
+func covertServerOptionsToMetricsServerOptions(serverOptions *options.ServerOption) []string {
 	var portedArgs = make([]string, 0)
 	if serverOptions == nil {
 		return portedArgs
@@ -91,11 +92,10 @@ func covertServerOptionsToMetricsServerOptions(serverOptions *options.ServerOpti
 	}
 	return portedArgs
 }
-func newMetricsAdpater(serverOptions *options.ServerOption, config *rest.Config, clusterStateCache clusterstatecache.Cache) *MetricsAdpater {
+func newMetricsAdpater(serverOptions *options.ServerOption, config *rest.Config, clusterStateCache clusterstatecache.Cache) *MetricsAdapter {
 	klog.V(10).Infof("[newMetricsAdpater] Entered newMetricsAdpater()")
 
-	cmd := &MetricsAdpater{
-	}
+	cmd := &MetricsAdapter{}
 
 	cmd.OpenAPIConfig = genericapiserver.DefaultOpenAPIConfig(generatedopenapi.GetOpenAPIDefinitions, openapinamer.NewDefinitionNamer(apiserver.Scheme))
 	cmd.OpenAPIConfig.Info.Title = "MetricsAdpater"