feat: add option continueAndFail on step.onError

Author: FaniD

cmd/entrypoint/main.go

@@ -54,7 +54,8 @@ var (
 	breakpointOnFailure = flag.Bool("breakpoint_on_failure", false, "If specified, expect steps to not skip on failure")
 	debugBeforeStep     = flag.Bool("debug_before_step", false, "If specified, wait for a debugger to attach before executing the step")
 	onError             = flag.String("on_error", "", "Set to \"continue\" to ignore an error and continue when a container terminates with a non-zero exit code."+
-		" Set to \"stopAndFail\" to declare a failure with a step error and stop executing the rest of the steps.")
+		" Set to \"stopAndFail\" to declare a failure with a step error and stop executing the rest of the steps."+
+		" Set to \"continueAndFail\" to declare a failure with a step error after having executed the rest of the steps.")
 	stepMetadataDir        = flag.String("step_metadata_dir", "", "If specified, create directory to store the step metadata e.g. /tekton/steps/<step-name>/")
 	resultExtractionMethod = flag.String("result_from", entrypoint.ResultExtractionMethodTerminationMessage, "The method using which to extract results from tasks. Default is using the termination message.")
 )

docs/pipeline-api.md

@@ -1752,6 +1752,9 @@ IncludeParamsList
 <tbody><tr><td><p>&#34;continue&#34;</p></td>
 <td><p>Continue indicates continue executing the rest of the steps irrespective of the container exit code</p>
 </td>
+</tr><tr><td><p>&#34;continueAndFail&#34;</p></td>
+<td><p>ContinueAndFail indicates continue executing the rest of the steps irrespective of the container exit code and exit after all steps are completed with the first possible non-zero exit code of the step that has this OnErrorType</p>
+</td>
 </tr><tr><td><p>&#34;stopAndFail&#34;</p></td>
 <td><p>StopAndFail indicates exit the taskRun if the container exits with non-zero exit code</p>
 </td>
@@ -4602,7 +4605,7 @@ OnErrorType
 </td>
 <td>
 <p>OnError defines the exiting behavior of a container on error
-can be set to [ continue | stopAndFail ]</p>
+can be set to [ continue | stopAndFail | continueAndFail ]</p>
 </td>
 </tr>
 <tr>

docs/tasks.md

@@ -385,11 +385,15 @@ When a `step` in a `task` results in a failure, the rest of the steps in the `ta
 declared a failure. If you would like to ignore such step errors and continue executing the rest of the steps in
 the task, you can specify `onError` for such a `step`.
 
-`onError` can be set to either `continue` or `stopAndFail` as part of the step definition. If `onError` is
+`onError` can be set to `continue`, `stopAndFail` or `continueAndFail` as part of the step definition. If `onError` is
 set to `continue`, the entrypoint sets the original failed exit code of the [script](#running-scripts-within-steps)
 in the container terminated state. A `step` with `onError` set to `continue` does not fail the `taskRun` and continues
 executing the rest of the steps in a task.
 
+If `onError` is set to `continueAndFail`, the `taskRun` continues executing the rest of the steps in a task and if no
+following `step` fails, the `taskRun` fails eventually with the exit code of the step that had set `onError` with
+`continueAndFail`.
+
 To ignore a step error, set `onError` to `continue`:
 
 ```yaml

pkg/apis/pipeline/v1/container_types.go

@@ -127,7 +127,7 @@ type Step struct {
 	Workspaces []WorkspaceUsage `json:"workspaces,omitempty"`
 
 	// OnError defines the exiting behavior of a container on error
-	// can be set to [ continue | stopAndFail ]
+	// can be set to [ continue | stopAndFail | continueAndFail ]
 	OnError OnErrorType `json:"onError,omitempty"`
 	// Stores configuration for the stdout stream of the step.
 	// +optional
@@ -176,6 +176,8 @@ const (
 	StopAndFail OnErrorType = "stopAndFail"
 	// Continue indicates continue executing the rest of the steps irrespective of the container exit code
 	Continue OnErrorType = "continue"
+	// Continue indicates continue executing the rest of the steps irrespective of the container exit code and exit after all steps are completed with the first possible non-zero exit code of the step that has this OnErrorType
+	ContinueAndFail OnErrorType = "continueAndFail"
 )
 
 // StepOutputConfig stores configuration for a step output stream.

pkg/apis/pipeline/v1/openapi_generated.go

@@ -1531,7 +1531,7 @@
           "default": ""
         },
         "onError": {
-          "description": "OnError defines the exiting behavior of a container on error can be set to [ continue | stopAndFail ]",
+          "description": "OnError defines the exiting behavior of a container on error can be set to [ continue | stopAndFail | continueAndFail ]",
           "type": "string"
         },
         "params": {

pkg/apis/pipeline/v1/swagger.json

@@ -505,11 +505,11 @@ func validateStep(ctx context.Context, s Step, names sets.String) (errs *apis.Fi
 	}
 
 	if s.OnError != "" {
-		if !isParamRefs(string(s.OnError)) && s.OnError != Continue && s.OnError != StopAndFail {
+		if !isParamRefs(string(s.OnError)) && s.OnError != Continue && s.OnError != StopAndFail && s.OnError != ContinueAndFail {
 			errs = errs.Also(&apis.FieldError{
 				Message: fmt.Sprintf("invalid value: \"%v\"", s.OnError),
 				Paths:   []string{"onError"},
-				Details: "Task step onError must be either \"continue\" or \"stopAndFail\"",
+				Details: "Task step onError must be \"continue\", \"stopAndFail\" or \"continueAndFail\"",
 			})
 		}
 	}

pkg/apis/pipeline/v1/task_validation.go

@@ -2279,6 +2279,13 @@ func TestStepOnError(t *testing.T) {
 			Image:   "image",
 			Args:    []string{"arg"},
 		}},
+	}, {
+		name: "valid step - valid onError usage - set to continueAndFail",
+		steps: []v1.Step{{
+			OnError: v1.ContinueAndFail,
+			Image:   "image",
+			Args:    []string{"arg"},
+		}},
 	}, {
 		name: "valid step - valid onError usage - set to a task parameter",
 		params: []v1.ParamSpec{{
@@ -2300,7 +2307,7 @@ func TestStepOnError(t *testing.T) {
 		expectedError: &apis.FieldError{
 			Message: `invalid value: "onError"`,
 			Paths:   []string{"steps[0].onError"},
-			Details: `Task step onError must be either "continue" or "stopAndFail"`,
+			Details: `Task step onError must be "continue", "stopAndFail" or "continueAndFail"`,
 		},
 	}}
 	for _, tt := range tests {

pkg/apis/pipeline/v1/task_validation_test.go

@@ -46,9 +46,10 @@ import (
 
 // RFC3339 with millisecond
 const (
-	timeFormat      = "2006-01-02T15:04:05.000Z07:00"
-	ContinueOnError = "continue"
-	FailOnError     = "stopAndFail"
+	timeFormat             = "2006-01-02T15:04:05.000Z07:00"
+	ContinueOnError        = "continue"
+	StopAndFailOnError     = "stopAndFail"
+	ContinueAndFailOnError = "continueAndFail"
 )
 
 const (
@@ -143,6 +144,7 @@ type Entrypointer struct {
 	// OnError defines exiting behavior of the entrypoint
 	// set it to "stopAndFail" to indicate the entrypoint to exit the taskRun if the container exits with non zero exit code
 	// set it to "continue" to indicate the entrypoint to continue executing the rest of the steps irrespective of the container exit code
+	// set it to "continueAndFail" to indicate the entrypoint to continue executing the rest of the steps irrespective of the container exit code and exit the taskRun if the container exits with non zero exit code
 	OnError string
 	// StepMetadataDir is the directory for a step where the step related metadata can be stored
 	StepMetadataDir string
@@ -294,6 +296,16 @@ func (e Entrypointer) Go() error {
 		})
 		e.WritePostFile(e.PostFile, nil)
 		e.WriteExitCodeFile(e.StepMetadataDir, exitCode)
+	case e.OnError == ContinueAndFailOnError && errors.As(err, &ee):
+		// with continueAndFail on error and an ExitError, write non-zero exit code and a post file with the error
+		exitCode := strconv.Itoa(ee.ExitCode())
+		output = append(output, result.RunResult{
+			Key:        "ExitCode",
+			Value:      exitCode,
+			ResultType: result.InternalTektonResultType,
+		})
+		e.WritePostFile(e.PostFile, err)
+		e.WriteExitCodeFile(e.StepMetadataDir, exitCode)
 	case err == nil:
 		// if err is nil, write zero exit code and a post file
 		e.WritePostFile(e.PostFile, nil)

pkg/entrypoint/entrypointer.go

@@ -501,12 +501,24 @@ func TestEntrypointer_OnError(t *testing.T) {
 		runner:        &fakeExitErrorRunner{},
 		expectedError: true,
 		postFile:      "step-one",
-		onError:       FailOnError,
+		onError:       StopAndFailOnError,
 	}, {
 		desc:          "the step is exiting with 0, treat the step error (but there is none) as failure with onError set to stopAndFail",
 		runner:        &fakeRunner{},
 		postFile:      "step-one",
-		onError:       FailOnError,
+		onError:       StopAndFailOnError,
+		expectedError: false,
+	}, {
+		desc:          "the step is exiting with 1, treat the step error as failure with onError set to continueAndFail",
+		runner:        &fakeExitErrorRunner{},
+		postFile:      "step-one",
+		onError:       ContinueAndFailOnError,
+		expectedError: true,
+	}, {
+		desc:          "the step is exiting with 0, treat the step error (but there is none) as failure with onError set to continueAndFail",
+		runner:        &fakeRunner{},
+		postFile:      "step-one",
+		onError:       ContinueAndFailOnError,
 		expectedError: false,
 	}, {
 		desc:            "the step set debug before step, and before step breakpoint fail-continue",
@@ -566,14 +578,30 @@ func TestEntrypointer_OnError(t *testing.T) {
 				}
 			}
 
-			if c.onError == FailOnError {
+			if c.onError == StopAndFailOnError {
 				switch {
 				case fpw.wrote == nil:
 					t.Error("Wanted post file written, got nil")
 				case c.expectedError && *fpw.wrote != c.postFile+".err":
 					t.Errorf("Wrote post file %q, want %q", *fpw.wrote, c.postFile+".err")
 				}
 			}
+
+			if c.onError == ContinueAndFailOnError {
+				switch {
+				case fpw.wrote == nil:
+					t.Error("Wanted post file written, got nil")
+				case fpw.exitCodeFile == nil:
+					t.Error("Wanted exitCode file written, got nil")
+				case *fpw.exitCodeFile != "exitCode":
+					t.Errorf("Wrote exitCode file %q, want %q", *fpw.exitCodeFile, "exitCode")
+				case c.expectedError && *fpw.exitCode == "0":
+					t.Errorf("Wrote zero exit code but want non-zero when expecting an error")
+				case c.expectedError && *fpw.wrote != c.postFile+".err":
+					t.Errorf("Wrote post file %q, want %q", *fpw.wrote, c.postFile+".err")
+				}
+			}
+
 		})
 	}
 }

pkg/entrypoint/entrypointer_test.go

@@ -154,9 +154,9 @@ func orderContainers(ctx context.Context, commonExtraEntrypointArgs []string, st
 		if taskSpec != nil {
 			if taskSpec.Steps != nil && len(taskSpec.Steps) >= i+1 {
 				if taskSpec.Steps[i].OnError != "" {
-					if taskSpec.Steps[i].OnError != v1.Continue && taskSpec.Steps[i].OnError != v1.StopAndFail {
-						return nil, fmt.Errorf("task step onError must be either \"%s\" or \"%s\" but it is set to an invalid value \"%s\"",
-							v1.Continue, v1.StopAndFail, taskSpec.Steps[i].OnError)
+					if taskSpec.Steps[i].OnError != v1.Continue && taskSpec.Steps[i].OnError != v1.StopAndFail && taskSpec.Steps[i].OnError != v1.ContinueAndFail {
+						return nil, fmt.Errorf("task step onError must have one of the following values: \"%s\", \"%s\" or \"%s\". But it is set to an invalid value \"%s\"",
+							v1.Continue, v1.StopAndFail, v1.ContinueAndFail, taskSpec.Steps[i].OnError)
 					}
 					argsForEntrypoint = append(argsForEntrypoint, "-on_error", string(taskSpec.Steps[i].OnError))
 				}

pkg/pod/entrypoint.go

@@ -1,4 +1,4 @@
-/*
+/*/
 Copyright 2019 The Tekton Authors
 
 Licensed under the Apache License, Version 2.0 (the "License");
@@ -645,6 +645,10 @@ func TestEntryPointOnError(t *testing.T) {
 		Name:    "passing-step",
 		Image:   "step-2",
 		Command: []string{"cmd"},
+	}, {
+		Name:    "passing-step",
+		Image:   "step-3",
+		Command: []string{"cmd"},
 	}}
 
 	for _, tc := range []struct {
@@ -658,6 +662,8 @@ func TestEntryPointOnError(t *testing.T) {
 				OnError: v1.Continue,
 			}, {
 				OnError: v1.StopAndFail,
+			}, {
+				OnError: v1.ContinueAndFail,
 			}},
 		},
 		wantContainers: []corev1.Container{{
@@ -688,14 +694,27 @@ func TestEntryPointOnError(t *testing.T) {
 				"-entrypoint", "cmd", "--",
 			},
 			TerminationMessagePath: "/tekton/termination",
+		}, {
+			Name:    "passing-step",
+			Image:   "step-3",
+			Command: []string{entrypointBinary},
+			Args: []string{
+				"-wait_file", "/tekton/run/1/out",
+				"-post_file", "/tekton/run/2/out",
+				"-termination_path", "/tekton/termination",
+				"-step_metadata_dir", "/tekton/run/2/status",
+				"-on_error", "continueAndFail",
+				"-entrypoint", "cmd", "--",
+			},
+			TerminationMessagePath: "/tekton/termination",
 		}},
 	}, {
 		taskSpec: v1.TaskSpec{
 			Steps: []v1.Step{{
 				OnError: "invalid-on-error",
 			}},
 		},
-		err: errors.New("task step onError must be either \"continue\" or \"stopAndFail\" but it is set to an invalid value \"invalid-on-error\""),
+		err: errors.New("task step onError must have one of the following values: \"continue\", \"stopAndFail\" or \"continueAndFail\". But it is set to an invalid value \"invalid-on-error\""),
 	}} {
 		t.Run(tc.desc, func(t *testing.T) {
 			got, err := orderContainers(context.Background(), []string{}, steps, &tc.taskSpec, nil, true, false)
@@ -711,7 +730,7 @@ func TestEntryPointOnError(t *testing.T) {
 					t.Fatalf("orderContainers: %v", err)
 				}
 				if d := cmp.Diff(tc.wantContainers, got); d != "" {
-					t.Errorf("Diff %s", diff.PrintWantGot(d))
+					t.Errorf("Edoooo eimaste\nDiff %s", diff.PrintWantGot(d))
 				}
 			}
 		})