Fix remaining reliability findings

2026-04-28 06:38:05 -04:00
parent b0041c5d18
commit 047d875fe6
8 changed files with 200 additions and 217 deletions
@@ -376,30 +376,49 @@ public sealed class GatewaySession
        await _closeLock.WaitAsync(cancellationToken).ConfigureAwait(false);
        try
        {
-            if (_state is SessionState.Closed)
+            try
            {
-                return new SessionCloseResult(SessionId, SessionState.Closed, AlreadyClosed: true);
+                if (_state is SessionState.Closed)
+                {
+                    return new SessionCloseResult(SessionId, SessionState.Closed, AlreadyClosed: true);
+                }
+
+                bool alreadyClosing = _closeStarted;
+                _closeStarted = true;
+                _state = SessionState.Closing;
+
+                if (_workerClient is not null)
+                {
+                    try
+                    {
+                        await _workerClient.ShutdownAsync(ShutdownTimeout, cancellationToken).ConfigureAwait(false);
+                    }
+                    catch (Exception exception)
+                    {
+                        try
+                        {
+                            _workerClient.Kill(reason);
+                        }
+                        catch (Exception killException)
+                        {
+                            throw new SessionCloseStartedException(
+                                $"Session {SessionId} close failed after worker shutdown started.",
+                                new AggregateException(exception, killException));
+                        }
+
+                        throw;
+                    }
+                }
+
+                _state = SessionState.Closed;
+                return new SessionCloseResult(SessionId, SessionState.Closed, alreadyClosing);
            }
-
-            bool alreadyClosing = _closeStarted;
-            _closeStarted = true;
-            _state = SessionState.Closing;
-
-            if (_workerClient is not null)
+            catch (Exception exception) when (exception is not SessionCloseStartedException)
            {
-                try
-                {
-                    await _workerClient.ShutdownAsync(ShutdownTimeout, cancellationToken).ConfigureAwait(false);
-                }
-                catch
-                {
-                    _workerClient.Kill(reason);
-                    throw;
-                }
+                throw new SessionCloseStartedException(
+                    $"Session {SessionId} close failed after the close lock was acquired.",
+                    exception);
            }
-
-            _state = SessionState.Closed;
-            return new SessionCloseResult(SessionId, SessionState.Closed, alreadyClosing);
        }
        finally
        {