From 80085085d9968ddce22f5d5ad676db317cbd43f6 Mon Sep 17 00:00:00 2001
From: chengyongru <chengyongru.ai@gmail.com>
Date: Wed, 22 Jul 2026 15:13:18 +0800
Subject: [PATCH] fix(exec): retain failed owner session cleanup

---
 nanobot/agent/tools/exec_session.py    | 17 ++++++++++++++++-
 tests/tools/test_exec_session_tools.py | 23 +++++++++++++++++++++++
 2 files changed, 39 insertions(+), 1 deletion(-)

diff --git a/nanobot/agent/tools/exec_session.py b/nanobot/agent/tools/exec_session.py
index 2fad2e0b..c00c2fd2 100644
--- a/nanobot/agent/tools/exec_session.py
+++ b/nanobot/agent/tools/exec_session.py
@@ -341,10 +341,25 @@ class ExecSessionManager:
             for sid, s in list(self._sessions.items()):
                 if s.owner_session_key == owner_session_key:
                     victims.append(self._sessions.pop(sid))
-        await asyncio.gather(
+        results = await asyncio.gather(
             *(s.kill() for s in victims),
             return_exceptions=True,
         )
+        failures = [
+            (session, result)
+            for session, result in zip(victims, results, strict=True)
+            if isinstance(result, BaseException)
+        ]
+        if failures:
+            async with self._lock:
+                for session, _ in failures:
+                    self._sessions[session.session_id] = session
+            if len(failures) == 1:
+                raise failures[0][1]
+            raise BaseExceptionGroup(
+                "failed to terminate exec sessions by owner",
+                [result for _, result in failures],
+            )
         return len(victims)
 
     async def _cleanup_locked(self) -> None:
diff --git a/tests/tools/test_exec_session_tools.py b/tests/tools/test_exec_session_tools.py
index ca7aafad..148baf5b 100644
--- a/tests/tools/test_exec_session_tools.py
+++ b/tests/tools/test_exec_session_tools.py
@@ -679,6 +679,29 @@ def test_terminate_by_owner_returns_zero_for_no_match(tmp_path):
     asyncio.run(run())
 
 
+def test_terminate_by_owner_retains_failed_sessions():
+    async def run() -> None:
+        manager = ExecSessionManager()
+        session = SimpleNamespace(
+            session_id="failed",
+            owner_session_key="cli:a",
+            kill=AsyncMock(side_effect=OSError("termination failed")),
+        )
+        manager._sessions[session.session_id] = session
+
+        with pytest.raises(OSError, match="termination failed"):
+            await manager.terminate_by_owner("cli:a")
+
+        assert manager._sessions == {session.session_id: session}
+        session.kill.assert_awaited_once()
+
+        session.kill.side_effect = None
+        assert await manager.terminate_by_owner("cli:a") == 1
+        assert manager._sessions == {}
+
+    asyncio.run(run())
+
+
 def test_terminate_by_owner_skips_sessions_without_owner_key(tmp_path):
     async def run() -> None:
         manager = ExecSessionManager()