fix(host_agent): disks behind a RAID controller each keep their own name (#5806)

smartctl reports every disk behind a PERC/MegaRAID controller as /dev/bus/0
with type megaraid,N; the helper named a drive by its path alone, so all of
VDNT-Docker02's and Docker03's disks read as one drive "0", each overwriting
the last. The slot is now part of the name (bus0-megaraid3), and Steward gives
a repeated name an ordinal so two drives never share a series. Agent 1.20.1.

fix(ansible): Steward stopping mid-run logs no traceback (#5749)

The run's done-callback called exception() on a cancelled task, which raises;
it now logs a run cut short by shutdown at INFO. A finished run's in-memory
state is dropped by a timer rather than a sleeping task, which shutdown found
still pending.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/plugins/host_agent/test_drive_health.py

@@ -80,6 +80,38 @@ def test_a_drive_with_no_smart_and_one_that_can_t_be_opened_say_so(maint):
     assert maint.drive_summary("/dev/sdd", None, None)["error"] == "smartctl couldn't read it"
 
 
+@pytest.mark.parametrize("name, kind, device", [
+    ("/dev/sda", "sat", "sda"),
+    ("/dev/nvme0", "nvme", "nvme0"),
+    ("/dev/bus/0", "megaraid,3", "bus0-megaraid3"),     # VDNT-Docker02's PERC (#5806)
+    ("/dev/bus/0", "sat+megaraid,12", "bus0-megaraid12"),
+    ("/dev/sg1", "cciss,0", "sg1-cciss0"),
+    ("/dev/sdc", None, "sdc"),
+])
+def test_each_drive_has_a_name_of_its_own(maint, name, kind, device):
+    assert maint.drive_name(name, kind) == device
+
+
+def test_disks_behind_one_controller_stay_apart(maint, monkeypatch):
+    def fake(args):
+        if args == ["--scan-open"]:
+            return 0, {"devices": [{"name": "/dev/bus/0", "type": "megaraid,0"},
+                                   {"name": "/dev/bus/0", "type": "megaraid,1"}]}
+        return 0, ATA
+    monkeypatch.setattr(maint, "smartctl_version", lambda: (7, 4))
+    monkeypatch.setattr(maint, "smartctl_json", fake)
+    assert [d["device"] for d in maint.do_drives()["drives"]] == ["bus0-megaraid0", "bus0-megaraid1"]
+
+
+def test_a_name_seen_twice_is_still_two_drives():
+    """What agent 1.20.0 sent from a PERC: every disk named "0"."""
+    report = {"drives": [{"device": "0", "passed": True}, {"device": "0", "passed": False},
+                         {"device": "0", "passed": True}]}
+    assert sorted(d["device"] for d in dh.drives_of(report)) == ["0", "0#2", "0#3"]
+    rows = {(m, r) for m, r, _ in dh.metric_rows(report, "box") if m == "drive_failing"}
+    assert len(rows) == 3
+
+
 def test_a_failing_disk_s_figures_still_count(maint):
     """Exit bit 3 says the disk is failing; its JSON is the point, not an error."""
     failing = {**ATA, "smart_status": {"passed": False}}

tests/test_ansible_run_end.py

@@ -0,0 +1,52 @@
+"""#5749: a redeploy during an Ansible run logged two CancelledError tracebacks
+from the run's done-callback and two "Task was destroyed but it is pending!"
+lines for its in-memory cleanup. Steward stopping mid-run is not an error: the
+next start marks the run interrupted."""
+import asyncio
+import logging
+
+from steward.ansible import executor
+from steward.ansible.runner import run_ended
+
+
+class _App:
+    logger = logging.getLogger("test.ansible.run_end")
+
+
+async def test_a_run_cut_short_by_shutdown_logs_no_error(caplog):
+    async def run():
+        await asyncio.sleep(60)
+
+    task = asyncio.create_task(run())
+    task.add_done_callback(run_ended(_App, "r1"))
+    await asyncio.sleep(0)
+    task.cancel()
+    with caplog.at_level(logging.INFO, logger="test.ansible.run_end"):
+        await asyncio.gather(task, return_exceptions=True)
+        await asyncio.sleep(0)                      # done-callbacks run on the next turn
+    assert [r.levelname for r in caplog.records] == ["INFO"]
+    assert "Ansible run r1 interrupted" in caplog.records[0].getMessage()
+
+
+async def test_a_run_that_raised_is_still_an_error(caplog):
+    async def run():
+        raise RuntimeError("ansible-playbook vanished")
+
+    task = asyncio.create_task(run())
+    task.add_done_callback(run_ended(_App, "r2"))
+    with caplog.at_level(logging.INFO, logger="test.ansible.run_end"):
+        await asyncio.gather(task, return_exceptions=True)
+        await asyncio.sleep(0)
+    assert [(r.levelname, r.getMessage()) for r in caplog.records] == [
+        ("ERROR", "Ansible run r2 raised: ansible-playbook vanished")]
+
+
+async def test_a_finished_run_s_state_is_dropped_by_a_timer_not_a_task(monkeypatch):
+    """A task left sleeping is what shutdown found pending; a timer isn't one."""
+    monkeypatch.setattr(executor, "_MEM_TTL_SECS", 0.01)
+    executor._run_lines["r3"] = ["ok"]
+    before = asyncio.all_tasks()
+    executor._schedule_cleanup("r3")
+    assert asyncio.all_tasks() == before
+    await asyncio.sleep(0.05)
+    assert "r3" not in executor._run_lines