fix(l4d2-web): server delete job now removes the DB row

The delete job ran l4d2ctl delete (host-side cleanup) but never removed the
Server row, so deleted servers kept appearing on /servers. Hard-delete the
row in the worker's success path and skip the post-op status refresh, since
the systemd unit is gone.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

l4d2web/services/job_worker.py

@@ -311,6 +311,15 @@ def run_job(job_id: int) -> None:
                 on_stderr=on_stderr,
                 should_cancel=should_cancel,
             )
+            # Host-side cleanup succeeded; remove the DB row so the server
+            # disappears from /servers. Status refresh is skipped — the
+            # systemd unit is gone and querying it would just log noise.
+            with session_scope() as db:
+                server = db.scalar(select(Server).where(Server.id == server_id))
+                if server is not None:
+                    db.delete(server)
+            finish_job(job_id, "succeeded", 0)
+            return
         else:
             raise ValueError(f"unknown job operation: {operation}")
 

l4d2web/tests/test_job_worker.py

@@ -260,6 +260,44 @@ def test_unexpected_exception_fails_job_with_exit_code_one(seeded_worker, monkey
     job = load_job(job_id)
     assert job.state == "failed"
     assert job.exit_code == 1
+    # Failed delete must keep the Server row.
+    with session_scope() as session:
+        assert session.scalar(select(Server).where(Server.id == ids.server_one)) is not None
+
+
+def test_successful_delete_removes_server_row(seeded_worker, monkeypatch) -> None:
+    app, ids = seeded_worker
+    job_id = add_job(ids.user, "delete", server_id=ids.server_one)
+
+    calls = []
+
+    def fake_delete(server_id, *, on_stdout=None, on_stderr=None, should_cancel=None):
+        del should_cancel
+        calls.append(server_id)
+        on_stdout("removing systemd unit")
+
+    def fake_status(name):  # pragma: no cover — would mean delete didn't skip refresh
+        pytest.fail(f"server_status must not be called after a successful delete (got {name!r})")
+
+    monkeypatch.setattr(l4d2_facade, "delete_server", fake_delete)
+    monkeypatch.setattr(l4d2_facade, "server_status", fake_status)
+
+    with app.app_context():
+        assert run_worker_once() is True
+
+    assert calls == [ids.server_one]
+    job = load_job(job_id)
+    assert job.state == "succeeded"
+    assert job.exit_code == 0
+    with session_scope() as session:
+        assert session.scalar(select(Server).where(Server.id == ids.server_one)) is None
+        # Sibling server is untouched.
+        assert session.scalar(select(Server).where(Server.id == ids.server_two)) is not None
+        # The delete job itself stays in the job log; outerjoin in views shows
+        # "-" for its (now-orphaned) server_id pointer.
+        finished_job = session.scalar(select(Job).where(Job.id == job_id))
+        assert finished_job is not None
+        assert finished_job.server_id == ids.server_one
 
 
 def test_same_server_jobs_do_not_overlap(seeded_worker, monkeypatch) -> None: