Merge pull request #128 from Azure/add-test-case-node10-job1-pod100

add test case with 10 virtual nodes and 100 pods

contrib/cmd/runkperf/commands/bench/node100_job1_pod3k.go

@@ -68,7 +68,7 @@ func benchNode100Job1Pod3KCaseRun(cliCtx *cli.Context) (*internaltypes.Benchmark
 	go func() {
 		defer wg.Done()
 
-		utils.RepeatJobWith3KPod(jobCtx, kubeCfgPath, "job1pod3k", jobInterval)
+		utils.RepeatJobWithPod(jobCtx, kubeCfgPath, "job1pod3k", "workload/3kpod.job.yaml", jobInterval)
 	}()
 
 	rgResult, derr := utils.DeployRunnerGroup(ctx,

contrib/cmd/runkperf/commands/bench/node10_job1_pod100.go

@@ -0,0 +1,96 @@
+package bench
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"time"
+
+	internaltypes "github.com/Azure/kperf/contrib/internal/types"
+	"github.com/Azure/kperf/contrib/internal/utils"
+
+	"github.com/urfave/cli"
+)
+
+var benchNode10Job1Pod100Case = cli.Command{
+	Name: "node10_job1_pod100",
+	Usage: `
+
+The test suite is to setup 10 virtual nodes and deploy one job with 100 pods on
+that nodes. It repeats to create and delete job. The load profile is fixed.
+	`,
+	Flags: append(
+		[]cli.Flag{
+			cli.IntFlag{
+				Name:  "total",
+				Usage: "Total requests per runner (There are 10 runners totally and runner's rate is 1)",
+				Value: 1000,
+			},
+		},
+		commonFlags...,
+	),
+	Action: func(cliCtx *cli.Context) error {
+		_, err := renderBenchmarkReportInterceptor(
+			addAPIServerCoresInfoInterceptor(benchNode10Job1Pod100CaseRun),
+		)(cliCtx)
+		return err
+	},
+}
+
+// benchNode10Job1Pod100CaseRun is for benchNode10Job1Pod100Case subcommand.
+func benchNode10Job1Pod100CaseRun(cliCtx *cli.Context) (*internaltypes.BenchmarkReport, error) {
+	ctx := context.Background()
+	kubeCfgPath := cliCtx.GlobalString("kubeconfig")
+
+	rgCfgFile, rgSpec, rgCfgFileDone, err := newLoadProfileFromEmbed(cliCtx,
+		"loadprofile/node10_job1_pod100.yaml")
+	if err != nil {
+		return nil, err
+	}
+	defer func() { _ = rgCfgFileDone() }()
+
+	vcDone, err := deployVirtualNodepool(ctx, cliCtx, "node10job1pod100",
+		100,
+		cliCtx.Int("cpu"),
+		cliCtx.Int("memory"),
+		cliCtx.Int("max-pods"),
+	)
+	if err != nil {
+		return nil, fmt.Errorf("failed to deploy virtual node: %w", err)
+	}
+	defer func() { _ = vcDone() }()
+
+	var wg sync.WaitGroup
+	wg.Add(1)
+
+	jobInterval := 5 * time.Second
+	jobCtx, jobCancel := context.WithCancel(ctx)
+	go func() {
+		defer wg.Done()
+
+		utils.RepeatJobWithPod(jobCtx, kubeCfgPath, "job1pod100", "workload/100pod.job.yaml", jobInterval)
+	}()
+
+	rgResult, derr := utils.DeployRunnerGroup(ctx,
+		cliCtx.GlobalString("kubeconfig"),
+		cliCtx.GlobalString("runner-image"),
+		rgCfgFile,
+		cliCtx.GlobalString("runner-flowcontrol"),
+		cliCtx.GlobalString("rg-affinity"),
+	)
+	jobCancel()
+	wg.Wait()
+
+	if derr != nil {
+		return nil, derr
+	}
+
+	return &internaltypes.BenchmarkReport{
+		Description: fmt.Sprintf(`
+Environment: 100 virtual nodes managed by kwok-controller,
+Workload: Deploy 1 job with 3,000 pods repeatedly. The parallelism is 100. The interval is %v`, jobInterval),
+		LoadSpec: *rgSpec,
+		Result:   *rgResult,
+		Info:     make(map[string]interface{}),
+	}, nil
+}

contrib/cmd/runkperf/commands/bench/root.go

@@ -52,6 +52,7 @@ var Command = cli.Command{
 		},
 	},
 	Subcommands: []cli.Command{
+		benchNode10Job1Pod100Case,
 		benchNode100Job1Pod3KCase,
 		benchNode100DeploymentNPod10KCase,
 	},

contrib/cmd/runkperf/commands/ekswarmup/command.go

@@ -117,7 +117,7 @@ var Command = cli.Command{
 		go func() {
 			defer wg.Done()
 
-			utils.RepeatJobWith3KPod(jobCtx, kubeCfgPath, "warmupjob", 5*time.Second)
+			utils.RepeatJobWithPod(jobCtx, kubeCfgPath, "warmupjob", "workload/3kpod.job.yaml", 5*time.Second)
 		}()
 
 		_, derr := utils.DeployRunnerGroup(ctx,

contrib/internal/manifests/loadprofile/node10_job1_pod100.yaml

@@ -0,0 +1,27 @@
+count: 1
+loadProfile:
+  version: 1
+  description: "node10-job1-pod100"
+  spec:
+    rate: 10
+    total: 1000
+    conns: 10
+    client: 10
+    contentType: json
+    disableHTTP2: false
+    maxRetries: 0
+    requests:
+      - staleList:
+          version: v1
+          resource: pods
+        shares: 1000 # chance 1000 / (1000 + 100 + 100)
+      - quorumList:
+          version: v1
+          resource: pods
+          limit: 1000
+        shares: 100 # chance 100 / (1000 + 100 + 100)
+      - quorumList:
+          version: v1
+          resource: events
+          limit: 1000
+        shares: 100 # chance 100 / (1000 + 100 + 100)

contrib/internal/manifests/workload/100pod.job.yaml

@@ -0,0 +1,31 @@
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: batchjobs
+spec:
+  completions: 100
+  parallelism: 10
+  template:
+    metadata:
+      labels:
+        app: fake-pod
+    spec:
+      restartPolicy: Never
+      affinity:
+        nodeAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            nodeSelectorTerms:
+            - matchExpressions:
+              - key: type
+                operator: In
+                values:
+                - kperf-virtualnodes
+      # A taints was added to an automatically created Node.
+      # You can remove taints of Node or add this tolerations.
+      tolerations:
+      - key: "kperf.io/nodepool"
+        operator: "Exists"
+        effect: "NoSchedule"
+      containers:
+      - name: fake-container
+        image: fake-image

contrib/internal/utils/utils.go

@@ -44,12 +44,10 @@ var (
 	EKSRunnerNodepoolInstanceType = "m4.4xlarge"
 )
 
-// RepeatJobWith3KPod repeats to deploy 3k pods.
-func RepeatJobWith3KPod(ctx context.Context, kubeCfgPath string, namespace string, internal time.Duration) {
+// RepeatJobWithPod repeats to deploy 3k pods.
+func RepeatJobWithPod(ctx context.Context, kubeCfgPath string, namespace string, target string, internal time.Duration) {
 	klog.V(0).Info("Repeat to create job with 3k pods")
 
-	target := "workload/3kpod.job.yaml"
-
 	data, err := manifests.FS.ReadFile(target)
 	if err != nil {
 		panic(fmt.Errorf("unexpected error when read %s from embed memory: %v",

examples/node10_job1_pod100.yaml

@@ -0,0 +1,25 @@
+  version: 1
+  description: "node10-job1-pod100"
+  spec:
+    rate: 10
+    total: 1000
+    conns: 10
+    client: 10
+    contentType: json
+    disableHTTP2: false
+    maxRetries: 0
+    requests:
+      - staleList:
+          version: v1
+          resource: pods
+        shares: 1000 # chance 1000 / (1000 + 100 + 100)
+      - quorumList:
+          version: v1
+          resource: pods
+          limit: 1000
+        shares: 100 # chance 100 / (1000 + 100 + 100)
+      - quorumList:
+          version: v1
+          resource: events
+          limit: 1000
+        shares: 100 # chance 100 / (1000 + 100 + 100)