Move failFast flag to DAG template spec

argoproj · sarabala1979 · Jul 1, 2019 · Jun 21, 2019 · Jun 21, 2019 · Jun 25, 2019
commit fb7f06cb900096afcaddb103b03b58eaa21faec8
diff --git a/api/openapi-spec/swagger.json b/api/openapi-spec/swagger.json
@@ -235,6 +235,10 @@
  "tasks"
  ],
  "properties": {
+ "failFast": {
+ "description": "This flag is for DAG logic. The DAG logic has a built-in \"fail fast\" feature to stop scheduling new steps, as soon as it detects that one of the DAG nodes is failed. Then it waits until all DAG nodes are completed before failing the DAG itself. The FailFast flag default is true, if set to false, it will allow a DAG to run all branches of the DAG to completion (either success or failure), regardless of the failed outcomes of branches in the DAG. More info and example about this feature at https://github.com/argoproj/argo/issues/1442",
+ "type": "boolean"
+ },
  "target": {
  "description": "Target are one or more names of targets to execute in a DAG",
  "type": "string"
@@ -1160,10 +1164,6 @@
  "description": "Entrypoint is a template reference to the starting point of the workflow",
  "type": "string"
  },
- "failFast": {
- "description": "This flag is for DAG logic. The DAG logic has a built-in \"fail fast\" feature to stop scheduling new steps, as soon as it detects that one of the DAG nodes is failed. Then it waits until all DAG nodes are completed before failing the DAG itself. The FailFast flag default is true, if set to false, it will allow a DAG to run all branches of the DAG to completion (either success or failure), regardless of the failed outcomes of branches in the DAG. More info and example about this feature at https://github.com/argoproj/argo/issues/1442",
- "type": "boolean"
- },
  "hostAliases": {
  "description": "HostAliases is an optional list of hosts and IPs that will be injected into the pod spec",
  "type": "array",

diff --git a/pkg/apis/workflow/v1alpha1/openapi_generated.go b/pkg/apis/workflow/v1alpha1/openapi_generated.go
diff --git a/pkg/apis/workflow/v1alpha1/types.go b/pkg/apis/workflow/v1alpha1/types.go
@@ -167,14 +167,6 @@ type WorkflowSpec struct {
 
  // HostAliases is an optional list of hosts and IPs that will be injected into the pod spec
  HostAliases []apiv1.HostAlias `json:"hostAliases,omitempty"`
-
- // This flag is for DAG logic. The DAG logic has a built-in "fail fast" feature to stop scheduling new steps,
- // as soon as it detects that one of the DAG nodes is failed. Then it waits until all DAG nodes are completed
- // before failing the DAG itself.
- // The FailFast flag default is true, if set to false, it will allow a DAG to run all branches of the DAG to
- // completion (either success or failure), regardless of the failed outcomes of branches in the DAG.
- // More info and example about this feature at https://github.com/argoproj/argo/issues/1442
- FailFast *bool `json:"failFast,omitempty"`
 }
 
 // Template is a reusable and composable unit of execution in a workflow
@@ -899,6 +891,14 @@ type DAGTemplate struct {
 
  // Tasks are a list of DAG tasks
  Tasks []DAGTask `json:"tasks"`
+
+ // This flag is for DAG logic. The DAG logic has a built-in "fail fast" feature to stop scheduling new steps,
+ // as soon as it detects that one of the DAG nodes is failed. Then it waits until all DAG nodes are completed
+ // before failing the DAG itself.
+ // The FailFast flag default is true, if set to false, it will allow a DAG to run all branches of the DAG to
+ // completion (either success or failure), regardless of the failed outcomes of branches in the DAG.
+ // More info and example about this feature at https://github.com/argoproj/argo/issues/1442
+ FailFast *bool `json:"failFast,omitempty"`
 }
 
 // DAGTask represents a node in the graph during DAG execution

diff --git a/pkg/apis/workflow/v1alpha1/zz_generated.deepcopy.go b/pkg/apis/workflow/v1alpha1/zz_generated.deepcopy.go
diff --git a/test/e2e/functional/dag-disable-failFast.yaml b/test/e2e/functional/dag-disable-failFast.yaml
@@ -3,7 +3,6 @@ kind: Workflow
 metadata:
  generateName: dag-primay-branch-
 spec:
- failFast: false
  entrypoint: statis
  templates:
  - name: a
@@ -32,6 +31,7 @@ spec:
  args: ["hello world"]
  - name: statis
  dag:
+ failFast: false
  tasks:
  - name: A
  template: a

diff --git a/workflow/controller/dag.go b/workflow/controller/dag.go
@@ -89,25 +89,27 @@ func (d *dagContext) assessDAGPhase(targetTasks []string, nodes map[string]wfv1.
  }
 
  if unsuccessfulPhase != "" {
- // If failFast set to false, we should return Running to continue this workflow for other DAG branch
- if d.wf.Spec.FailFast != nil && !*d.wf.Spec.FailFast {
- tmpOverAllFinished := true
- // If all the nodes have finished, we should mark the failed node to finish overall workflow
- // So we should check all the targetTasks have finished
- for _, tmpDepName := range targetTasks {
- tmpDepNode := d.getTaskNode(tmpDepName)
- if tmpDepNode == nil {
- tmpOverAllFinished = false
- break
+ if d.tmpl != nil && d.tmpl.DAG != nil {
+ // If failFast set to false, we should return Running to continue this workflow for other DAG branch
+ if d.tmpl.DAG.FailFast != nil && !*d.tmpl.DAG.FailFast {
+ tmpOverAllFinished := true
+ // If all the nodes have finished, we should mark the failed node to finish overall workflow
+ // So we should check all the targetTasks have finished
+ for _, tmpDepName := range targetTasks {
+ tmpDepNode := d.getTaskNode(tmpDepName)
+ if tmpDepNode == nil {
+ tmpOverAllFinished = false
+ break
+ }
+ if tmpDepNode.Type == wfv1.NodeTypeRetry && hasMoreRetries(tmpDepNode, d.wf) {
+ tmpOverAllFinished = false
+ break
+ }
  }
- if tmpDepNode.Type == wfv1.NodeTypeRetry && hasMoreRetries(tmpDepNode, d.wf) {
- tmpOverAllFinished = false
- break
+ if !tmpOverAllFinished {
+ return wfv1.NodeRunning
  }
  }
- if !tmpOverAllFinished {
- return wfv1.NodeRunning
- }
  }
 
  // if we were unsuccessful, we can return *only* if all retry nodes have ben exhausted.

diff --git a/workflow/controller/dag_test.go b/workflow/controller/dag_test.go
@@ -32,3 +32,11 @@ func TestDagRetryExhaustedXfail(t *testing.T) {
  woc.operate()
  assert.Equal(t, string(wfv1.NodeFailed), string(woc.wf.Status.Phase))
 }
+
+// TestDagDisableFailFast test disable fail fast function
+func TestDagDisableFailFast(t *testing.T) {
+ wf := test.LoadTestWorkflow("testdata/dag-disable-fail-fast.yaml")
+ woc := newWoc(*wf)
+ woc.operate()
+ assert.Equal(t, string(wfv1.NodeFailed), string(woc.wf.Status.Phase))
+}
diff --git a/workflow/controller/testdata/dag-disable-fail-fast.yaml b/workflow/controller/testdata/dag-disable-fail-fast.yaml
@@ -0,0 +1,211 @@
+apiVersion: argoproj.io/v1alpha1
+kind: Workflow
+metadata:
+ creationTimestamp: 2019-06-26T09:11:58Z
+ generateName: dag-disable-fail-fast-
+ generation: 1
+ labels:
+ workflows.argoproj.io/completed: "true"
+ workflows.argoproj.io/phase: Failed
+ name: dag-disable-fail-fast-r6xdc
+ namespace: default
+ resourceVersion: "15772210"
+ selfLink: /apis/argoproj.io/v1alpha1/namespaces/default/workflows/dag-disable-fail-fast-r6xdc
+ uid: 734516b5-97f2-11e9-9fea-00163e00cf4e
+spec:
+ arguments: {}
+ entrypoint: statis
+ templates:
+ - container:
+ args:
+ - hello world
+ command:
+ - cowsay
+ image: docker/whalesay:latest
+ name: ""
+ resources: {}
+ inputs: {}
+ metadata: {}
+ name: a
+ outputs: {}
+ - container:
+ args:
+ - sleep 30; echo haha
+ command:
+ - sh
+ - -c
+ image: alpine:latest
+ name: ""
+ resources: {}
+ inputs: {}
+ metadata: {}
+ name: b
+ outputs: {}
+ retryStrategy:
+ limit: 2
+ - container:
+ args:
+ - echo intentional failure; exit 2
+ command:
+ - sh
+ - -c
+ image: alpine:latest
+ name: ""
+ resources: {}
+ inputs: {}
+ metadata: {}
+ name: c
+ outputs: {}
+ retryStrategy:
+ limit: 1
+ - container:
+ args:
+ - hello world
+ command:
+ - cowsay
+ image: docker/whalesay:latest
+ name: ""
+ resources: {}
+ inputs: {}
+ metadata: {}
+ name: d
+ outputs: {}
+ - dag:
+ failFast: false
+ tasks:
+ - arguments: {}
+ name: A
+ template: a
+ - arguments: {}
+ dependencies:
+ - A
+ name: B
+ template: b
+ - arguments: {}
+ dependencies:
+ - A
+ name: C
+ template: c
+ - arguments: {}
+ dependencies:
+ - B
+ name: D
+ template: d
+ - arguments: {}
+ dependencies:
+ - D
+ name: E
+ template: d
+ inputs: {}
+ metadata: {}
+ name: statis
+ outputs: {}
+status:
+ finishedAt: 2019-06-26T09:12:46Z
+ nodes:
+ dag-disable-fail-fast-r6xdc:
+ children:
+ - dag-disable-fail-fast-r6xdc-3928436299
+ displayName: dag-disable-fail-fast-r6xdc
+ finishedAt: 2019-06-26T09:12:46Z
+ id: dag-disable-fail-fast-r6xdc
+ name: dag-disable-fail-fast-r6xdc
+ phase: Failed
+ startedAt: 2019-06-26T09:11:58Z
+ templateName: statis
+ type: DAG
+ dag-disable-fail-fast-r6xdc-3256495944:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ displayName: C(0)
+ finishedAt: 2019-06-26T09:12:08Z
+ id: dag-disable-fail-fast-r6xdc-3256495944
+ message: failed with exit code 2
+ name: dag-disable-fail-fast-r6xdc.C(0)
+ phase: Failed
+ startedAt: 2019-06-26T09:12:03Z
+ templateName: c
+ type: Pod
+ dag-disable-fail-fast-r6xdc-3457680277:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ displayName: C(1)
+ finishedAt: 2019-06-26T09:12:12Z
+ id: dag-disable-fail-fast-r6xdc-3457680277
+ message: failed with exit code 2
+ name: dag-disable-fail-fast-r6xdc.C(1)
+ phase: Failed
+ startedAt: 2019-06-26T09:12:09Z
+ templateName: c
+ type: Pod
+ dag-disable-fail-fast-r6xdc-3928436299:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ children:
+ - dag-disable-fail-fast-r6xdc-3945213918
+ - dag-disable-fail-fast-r6xdc-3961991537
+ displayName: A
+ finishedAt: 2019-06-26T09:12:02Z
+ id: dag-disable-fail-fast-r6xdc-3928436299
+ name: dag-disable-fail-fast-r6xdc.A
+ phase: Succeeded
+ startedAt: 2019-06-26T09:11:58Z
+ templateName: a
+ type: Pod
+ dag-disable-fail-fast-r6xdc-3945213918:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ children:
+ - dag-disable-fail-fast-r6xdc-4286504589
+ displayName: B
+ finishedAt: 2019-06-26T09:12:36Z
+ id: dag-disable-fail-fast-r6xdc-3945213918
+ name: dag-disable-fail-fast-r6xdc.B
+ phase: Succeeded
+ startedAt: 2019-06-26T09:12:03Z
+ type: Retry
+ dag-disable-fail-fast-r6xdc-3961991537:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ children:
+ - dag-disable-fail-fast-r6xdc-3256495944
+ - dag-disable-fail-fast-r6xdc-3457680277
+ displayName: C
+ finishedAt: 2019-06-26T09:12:13Z
+ id: dag-disable-fail-fast-r6xdc-3961991537
+ message: No more retries left
+ name: dag-disable-fail-fast-r6xdc.C
+ phase: Failed
+ startedAt: 2019-06-26T09:12:03Z
+ type: Retry
+ dag-disable-fail-fast-r6xdc-3978769156:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ children:
+ - dag-disable-fail-fast-r6xdc-3995546775
+ displayName: D
+ finishedAt: 2019-06-26T09:12:41Z
+ id: dag-disable-fail-fast-r6xdc-3978769156
+ name: dag-disable-fail-fast-r6xdc.D
+ phase: Succeeded
+ startedAt: 2019-06-26T09:12:37Z
+ templateName: d
+ type: Pod
+ dag-disable-fail-fast-r6xdc-3995546775:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ displayName: E
+ finishedAt: 2019-06-26T09:12:45Z
+ id: dag-disable-fail-fast-r6xdc-3995546775
+ name: dag-disable-fail-fast-r6xdc.E
+ phase: Succeeded
+ startedAt: 2019-06-26T09:12:42Z
+ templateName: d
+ type: Pod
+ dag-disable-fail-fast-r6xdc-4286504589:
+ boundaryID: dag-disable-fail-fast-r6xdc
+ children:
+ - dag-disable-fail-fast-r6xdc-3978769156
+ displayName: B(0)
+ finishedAt: 2019-06-26T09:12:36Z
+ id: dag-disable-fail-fast-r6xdc-4286504589
+ name: dag-disable-fail-fast-r6xdc.B(0)
+ phase: Succeeded
+ startedAt: 2019-06-26T09:12:03Z
+ templateName: b
+ type: Pod
+ phase: Failed
+ startedAt: 2019-06-26T09:11:58Z