fix(e2e): use the unified job commands

cmd/ob/ops.go

@@ -64,7 +64,7 @@ func addOpsCommands(root *cobra.Command, g *globalFlags) {
 	proxyCmd.AddCommand(proxyApplyCmd)
 	root.AddCommand(proxyCmd)
 
-	// schedule apply | list | history | logs — see schedule.go.
+	// schedule apply | list — see schedule.go; job run | history | logs — see job.go.
 	addScheduleCommands(root, g)
 
 	// secrets list | edit | push

e2e/server_execution_test.go

@@ -71,7 +71,7 @@ workloads:
 	// A legacy compose-run container can survive a crash without durable labels.
 	// The first durable activation must reclaim this stopped, owned container.
 	s.run(t, "docker compose -p "+name+" --project-directory "+root+"/current -f "+root+"/current/compose.yaml run --no-deps --name "+name+"-refresh-1 refresh true")
-	if out, err := s.ob(t, dir, "schedule", "run", "refresh", "--input", "SOURCE=custom", "--wait"); err == nil {
+	if out, err := s.ob(t, dir, "job", "run", "refresh", "--input", "SOURCE=custom", "--wait"); err == nil {
 		t.Fatalf("index should fail before allow marker: %s", out)
 	}
 	list := s.mustOb(t, dir, "execution", "list", "--output", "json")
@@ -114,15 +114,15 @@ workloads:
 	// Interrupted work is retained independently of the host journal. A new
 	// failed execution can be inspected after its runner and notifier exit.
 	s.run(t, "rm -f "+base+"/data/allow")
-	_, _ = s.ob(t, dir, "schedule", "run", "refresh", "--wait")
+	_, _ = s.ob(t, dir, "job", "run", "refresh", "--wait")
 	pins := s.run(t, "/usr/bin/python3 "+root+"/schedule/execution-v1.py pins "+root)
 	if strings.TrimSpace(pins) == "" {
 		t.Fatal("failed execution lost its durable release reference")
 	}
 	// A killed activation must be inspectable and resume only its interrupted
 	// step. Concurrent resume and deployment coordination must refuse live work.
 	s.run(t, "touch "+base+"/data/hold")
-	s.mustOb(t, dir, "schedule", "run", "refresh")
+	s.mustOb(t, dir, "job", "run", "refresh")
 	deadline := time.Now().Add(15 * time.Second)
 	for s.try(t, "test -f "+base+"/data/started") != nil {
 		if time.Now().After(deadline) {

e2e/server_test.go

@@ -238,7 +238,7 @@ ExecStart=/usr/bin/docker compose -p observer -f /var/lib/ob/observer/current/co
 
 		// The notifier wrote a record for that run; a hand-started unit has no
 		// TRIGGER_UNIT, so it is recorded as a operator activation.
-		history := s.mustOb(t, dir, "schedule", "history", "chore", "--output", "json")
+		history := s.mustOb(t, dir, "job", "history", "chore", "--output", "json")
 		for _, want := range []string{`"outcome": "success"`, `"trigger": "manual"`, `"attempts": 1`} {
 			if !strings.Contains(history, want) {
 				t.Fatalf("run history is missing %q:\n%s", want, history)
@@ -248,7 +248,7 @@ ExecStart=/usr/bin/docker compose -p observer -f /var/lib/ob/observer/current/co
 		// One failure, one sleep, one success: the record counts both attempts.
 		s.run(t, "rm -rf /tmp/onebox-e2e-retry && mkdir -p /tmp/onebox-e2e-retry")
 		s.run(t, "systemctl start ob-observer-retry--chore.service")
-		retry := s.mustOb(t, dir, "schedule", "history", "retry-chore", "--output", "json")
+		retry := s.mustOb(t, dir, "job", "history", "retry-chore", "--output", "json")
 		for _, want := range []string{`"outcome": "success"`, `"attempts": 2`} {
 			if !strings.Contains(retry, want) {
 				t.Fatalf("retry history is missing %q:\n%s", want, retry)
@@ -257,13 +257,13 @@ ExecStart=/usr/bin/docker compose -p observer -f /var/lib/ob/observer/current/co
 
 		// A operator run with an input override reaches the container as its
 		// environment, is journaled with the operator, and shows up as such.
-		manual := s.mustOb(t, dir, "schedule", "run", "input-chore", "--input", "GREETING=hello", "--wait", "--output", "json")
+		manual := s.mustOb(t, dir, "job", "run", "input-chore", "--input", "GREETING=hello", "--wait", "--output", "json")
 		for _, want := range []string{`"GREETING": "hello"`, `"outcome": "success"`, `"trigger": "manual"`} {
 			if !strings.Contains(manual, want) {
 				t.Fatalf("operator run result is missing %q:\n%s", want, manual)
 			}
 		}
-		logs := s.mustOb(t, dir, "schedule", "logs", "input-chore")
+		logs := s.mustOb(t, dir, "job", "logs", "input-chore")
 		if !strings.Contains(logs, "greeting=hello") {
 			t.Fatalf("run logs do not show the override:\n%s", logs)
 		}
@@ -312,7 +312,7 @@ HTTPServer(("127.0.0.1", 18080), Handler).handle_request()
 			t.Fatalf("timed-out scheduled run result = %q, want timeout", result)
 		}
 		// The runner was killed mid-run; ExecStopPost still wrote the record.
-		if timedOut := s.mustOb(t, dir, "schedule", "history", "timeout-chore", "--output", "json"); !strings.Contains(timedOut, `"outcome": "timeout"`) {
+		if timedOut := s.mustOb(t, dir, "job", "history", "timeout-chore", "--output", "json"); !strings.Contains(timedOut, `"outcome": "timeout"`) {
 			t.Fatalf("timeout was not recorded:\n%s", timedOut)
 		}
 		out, err := s.ob(t, dir, "status")
@@ -334,7 +334,7 @@ HTTPServer(("127.0.0.1", 18080), Handler).handle_request()
 		// clears the failure from `ob status`; only a later successful run
 		// does. A operator run with the input that makes the job finish in time
 		// is that run, and it must leave status green for the steps after.
-		s.mustOb(t, dir, "schedule", "run", "timeout-chore", "--input", "SLEEP=0", "--wait")
+		s.mustOb(t, dir, "job", "run", "timeout-chore", "--input", "SLEEP=0", "--wait")
 		if cleared := s.mustOb(t, dir, "status"); !strings.Contains(cleared, "schedule timeout-chore active") {
 			t.Fatalf("a successful operator run did not clear the recorded timeout:\n%s", cleared)
 		}