winrm - Handle task timeout (#82784) (#82862)

When using winrm over HTTP with message encryption enabled and a task
has timed out the connection plugin will fail to cleanup the WinRM
command. This will change that exception into a warning as a timeout is
already an exception event and a failure to clean the operation should
not override the timeout error shown.

(cherry picked from commit 8aecd1f9b2)

changelogs/fragments/winrm-task-timeout.yml

@@ -0,0 +1,2 @@
+bugfixes:
+- winrm - Do not raise another exception during cleanup when a task is timed out - https://github.com/ansible/ansible/issues/81095

lib/ansible/plugins/connection/winrm.py

@@ -177,7 +177,7 @@ from ansible.utils.display import Display
 try:
     import winrm
     from winrm import Response
-    from winrm.exceptions import WinRMError, WinRMOperationTimeoutError
+    from winrm.exceptions import WinRMError, WinRMOperationTimeoutError, WinRMTransportError
     from winrm.protocol import Protocol
     import requests.exceptions
     HAS_WINRM = True
@@ -582,7 +582,19 @@ class Connection(ConnectionBase):
             raise AnsibleConnectionFailure('winrm connection error: %s' % to_native(exc))
         finally:
             if command_id:
-                self.protocol.cleanup_command(self.shell_id, command_id)
+                # Due to a bug in how pywinrm works with message encryption we
+                # ignore a 400 error which can occur when a task timeout is
+                # set and the code tries to clean up the command. This happens
+                # as the cleanup msg is sent over a new socket but still uses
+                # the already encrypted payload bound to the other socket
+                # causing the server to reply with 400 Bad Request.
+                try:
+                    self.protocol.cleanup_command(self.shell_id, command_id)
+                except WinRMTransportError as e:
+                    if e.code != 400:
+                        raise
+
+                    display.warning("Failed to cleanup running WinRM command, resources might still be in use on the target server")
 
     def _connect(self):
 

test/integration/targets/connection_windows_ssh/tests.yml

@@ -30,3 +30,15 @@
       - win_ssh_async.rc == 0
       - win_ssh_async.stdout == "café\n"
       - win_ssh_async.stderr == ""
+
+  # Ensures the connection plugin can handle a timeout
+  # without raising another error.
+  - name: run command with timeout
+    win_shell: Start-Sleep -Seconds 10
+    timeout: 5
+    register: timeout_cmd
+    ignore_errors: true
+
+  - assert:
+      that:
+      - timeout_cmd.msg == 'The win_shell action failed to execute in the expected time frame (5) and was terminated'

test/integration/targets/connection_winrm/tests.yml

@@ -26,3 +26,15 @@
       that:
       - winrm_copy_empty is changed
       - winrm_copy_empty_actual.stat.size == 0
+
+  # Ensures the connection plugin can handle a timeout
+  # without raising another error.
+  - name: run command with timeout
+    win_shell: Start-Sleep -Seconds 10
+    timeout: 5
+    register: timeout_cmd
+    ignore_errors: true
+
+  - assert:
+      that:
+      - timeout_cmd.msg == 'The win_shell action failed to execute in the expected time frame (5) and was terminated'