Backport of Prevent kill_timeout greater than progress_deadline into …

…release/1.4.x (#17205) * test: fix TestJobEndpoint_Scale_BatchJob * Prevent kill_timeout greater than progress_deadline (#16761) * func: add validation for kill timeout smaller than progress dealine * style: add changelog * style: typo in changelog * style: remove refactored test * Update .changelog/16761.txt Co-authored-by: James Rasell <jrasell@users.noreply.github.com> * Update nomad/structs/structs.go Co-authored-by: James Rasell <jrasell@users.noreply.github.com> --------- Co-authored-by: James Rasell <jrasell@users.noreply.github.com> --------- Co-authored-by: Luiz Aoqui <luiz@hashicorp.com> Co-authored-by: Juana De La Cuesta <juanita.delacuestamorales@hashicorp.com> Co-authored-by: James Rasell <jrasell@users.noreply.github.com>
hashicorp · May 16, 2023 · c7ad698 · c7ad698
1 parent 6ad39fe
commit c7ad698
Show file tree

Hide file tree

Showing 3 changed files with 427 additions and 328 deletions.
diff --git a/.changelog/16761.txt b/.changelog/16761.txt
@@ -0,0 +1,3 @@
+```release-note:improvement
+core: Prevent `task.kill_timeout` being greater than `update.progress_deadline`
+```
diff --git a/nomad/structs/structs.go b/nomad/structs/structs.go
@@ -4972,6 +4972,12 @@ func (u *UpdateStrategy) IsEmpty() bool {
 		return true
 	}
 
+	// When the Job is transformed from api to struct, the Update Strategy block is
+	// copied into the existing task groups, the only things that are passed along
+	// are MaxParallel and Stagger, because they are enforced at job level.
+	// That is why checking if MaxParallel is zero is enough to know if the
+	// update block is empty.
+
 	return u.MaxParallel == 0
 }
 
@@ -6540,6 +6546,8 @@ func (tg *TaskGroup) Validate(j *Job) error {
 		mErr.Errors = append(mErr.Errors, outer)
 	}
 
+	isTypeService := j.Type == JobTypeService
+
 	// Validate the tasks
 	for _, task := range tg.Tasks {
 		// Validate the task does not reference undefined volume mounts
@@ -6559,6 +6567,15 @@ func (tg *TaskGroup) Validate(j *Job) error {
 			outer := fmt.Errorf("Task %s validation failed: %v", task.Name, err)
 			mErr.Errors = append(mErr.Errors, outer)
 		}
+
+		// Validate the group's Update Strategy does not conflict with the Task's kill_timeout for service type jobs
+		if isTypeService && tg.Update != nil {
+			if task.KillTimeout > tg.Update.ProgressDeadline {
+				mErr.Errors = append(mErr.Errors, fmt.Errorf("Task %s has a kill timout (%s) longer than the group's progress deadline (%s)",
+					task.Name, task.KillTimeout.String(), tg.Update.ProgressDeadline.String()))
+			}
+		}
+
 	}
 
 	return mErr.ErrorOrNil()