Merge pull request #3821 from hashicorp/f-restart-stanza-fail-default

Change the default mode for client side restarts to fail from delay

CHANGELOG.md

@@ -4,6 +4,9 @@ __BACKWARDS INCOMPATIBILITIES:__
  * discovery: Prevent absolute URLs in check paths. The documentation indicated
    that absolute URLs are not allowed, but it was not enforced. Absolute URLs
    in HTTP check paths will now fail to validate. [[GH-3685](https://github.com/hashicorp/nomad/issues/3685)]
+ * jobspec: The default values for restart policy have changed. Restart policy mode defaults to "fail" and the
+   attempts/time interval values have been changed to enable faster server side rescheduling. See
+   [restart stanza](https://www.nomadproject.io/docs/job-specification/restart.html) for more information.
 
 IMPROVEMENTS:
  * core: A set of features (Autopilot) has been added to allow for automatic operator-friendly management of Nomad servers. For more information about Autopilot, see the [Autopilot Guide](https://www.nomadproject.io/guides/cluster/autopilot.html). [[GH-3670](https://github.com/hashicorp/nomad/pull/3670)]

api/tasks.go

@@ -340,17 +340,17 @@ func (g *TaskGroup) Canonicalize(job *Job) {
 	switch *job.Type {
 	case "service", "system":
 		defaultRestartPolicy = &RestartPolicy{
-			Delay:    helper.TimeToPtr(15 * time.Second),
-			Attempts: helper.IntToPtr(2),
-			Interval: helper.TimeToPtr(1 * time.Minute),
-			Mode:     helper.StringToPtr("delay"),
+			Delay:    helper.TimeToPtr(structs.DefaultServiceJobRestartPolicy.Delay),
+			Attempts: helper.IntToPtr(structs.DefaultServiceJobRestartPolicy.Attempts),
+			Interval: helper.TimeToPtr(structs.DefaultServiceJobRestartPolicy.Interval),
+			Mode:     helper.StringToPtr(structs.DefaultServiceJobRestartPolicy.Mode),
 		}
 	default:
 		defaultRestartPolicy = &RestartPolicy{
-			Delay:    helper.TimeToPtr(15 * time.Second),
-			Attempts: helper.IntToPtr(15),
-			Interval: helper.TimeToPtr(7 * 24 * time.Hour),
-			Mode:     helper.StringToPtr("delay"),
+			Delay:    helper.TimeToPtr(structs.DefaultBatchJobRestartPolicy.Delay),
+			Attempts: helper.IntToPtr(structs.DefaultBatchJobRestartPolicy.Attempts),
+			Interval: helper.TimeToPtr(structs.DefaultBatchJobRestartPolicy.Interval),
+			Mode:     helper.StringToPtr(structs.DefaultBatchJobRestartPolicy.Mode),
 		}
 	}
 

command/init.go

@@ -183,18 +183,18 @@ job "example" {
     #
     restart {
       # The number of attempts to run the job within the specified interval.
-      attempts = 10
-      interval = "5m"
+      attempts = 2
+      interval = "15s"
 
       # The "delay" parameter specifies the duration to wait before restarting
       # a task after it has failed.
-      delay = "25s"
+      delay = "15s"
 
      # The "mode" parameter controls what happens when a task has restarted
      # "attempts" times within the interval. "delay" mode delays the next
      # restart until the next interval. "fail" mode does not restart the task
      # if "attempts" has been hit within the interval.
-      mode = "delay"
+      mode = "fail"
     }
 
     # The "ephemeral_disk" stanza instructs Nomad to utilize an ephemeral disk

nomad/structs/structs.go

@@ -2510,17 +2510,17 @@ func (d *DispatchPayloadConfig) Validate() error {
 }
 
 var (
-	defaultServiceJobRestartPolicy = RestartPolicy{
+	DefaultServiceJobRestartPolicy = RestartPolicy{
 		Delay:    15 * time.Second,
 		Attempts: 2,
-		Interval: 1 * time.Minute,
-		Mode:     RestartPolicyModeDelay,
+		Interval: 30 * time.Minute,
+		Mode:     RestartPolicyModeFail,
 	}
-	defaultBatchJobRestartPolicy = RestartPolicy{
+	DefaultBatchJobRestartPolicy = RestartPolicy{
 		Delay:    15 * time.Second,
-		Attempts: 15,
-		Interval: 7 * 24 * time.Hour,
-		Mode:     RestartPolicyModeDelay,
+		Attempts: 3,
+		Interval: 24 * time.Hour,
+		Mode:     RestartPolicyModeFail,
 	}
 )
 
@@ -2604,10 +2604,10 @@ func (r *RestartPolicy) Validate() error {
 func NewRestartPolicy(jobType string) *RestartPolicy {
 	switch jobType {
 	case JobTypeService, JobTypeSystem:
-		rp := defaultServiceJobRestartPolicy
+		rp := DefaultServiceJobRestartPolicy
 		return &rp
 	case JobTypeBatch:
-		rp := defaultBatchJobRestartPolicy
+		rp := DefaultBatchJobRestartPolicy
 		return &rp
 	}
 	return nil

website/source/api/json-jobs.html.md

@@ -91,10 +91,10 @@ Below is the JSON representation of the job outputted by `$ nomad init`:
                 "Leader": false
             }],
             "RestartPolicy": {
-                "Interval": 300000000000,
-                "Attempts": 10,
-                "Delay": 25000000000,
-                "Mode": "delay"
+                "Interval": 1800000000000,
+                "Attempts": 2,
+                "Delay": 15000000000,
+                "Mode": "fail"
             },
             "EphemeralDisk": {
                 "SizeMB": 300

website/source/docs/job-specification/restart.html.md

@@ -17,7 +17,8 @@ description: |-
   </tr>
 </table>
 
-The `restart` stanza configures a group's behavior on task failure.
+The `restart` stanza configures a group's behavior on task failure. Restarts
+happen on the client that is running the task.
 
 ```hcl
 job "docs" {
@@ -62,7 +63,7 @@ defaults by job type:
       attempts = 15
       delay    = "15s"
       interval = "168h"
-      mode     = "delay"
+      mode     = "fail"
     }
     ```
 
@@ -73,7 +74,7 @@ defaults by job type:
       interval = "1m"
       attempts = 2
       delay    = "15s"
-      mode     = "delay"
+      mode     = "fail"
     }
     ```