Fix prompt mismatch issue for ios (#47004)

* Fix prompt mismatch issue for ios

Fixes #40884 #44463

*  If the command prompt is matched check if data is
   still pending to be read from buffer.
*  This fix adds a new timer `buffer_read_timeout`
   which will be trigerred after command prompt
   is matched and data is attempted to be read from channel.
   If not data is present of channel the timer will expire
   and response we be returned to calling function.

* Fix unit test failure

* Update to make buffer timeout float

* Update doc and fix review comment

* Fix CI issues

* Update doc

* Fix review comments

* Update changelog

(cherry picked from commit 335a979f1d30d065504bf9222460f35356a2e0b6)

4 files changed, 111 insertions, 9 deletions

diff --git a/changelogs/fragments/ios-prompt-issue.yaml b/changelogs/fragments/ios-prompt-issue.yaml
new file mode 100644
index 0000000000..0c0fc632ac
--- /dev/null
+++ b/changelogs/fragments/ios-prompt-issue.yaml
@@ -0,0 +1,2 @@
+bugfixes:
+- Fix prompt mismatch issue for ios (https://github.com/ansible/ansible/issues/47004)
diff --git a/docs/docsite/rst/network/user_guide/network_debug_troubleshooting.rst b/docs/docsite/rst/network/user_guide/network_debug_troubleshooting.rst
index a68f00e9c7..3d734cd851 100644
--- a/docs/docsite/rst/network/user_guide/network_debug_troubleshooting.rst
+++ b/docs/docsite/rst/network/user_guide/network_debug_troubleshooting.rst
@@ -656,3 +656,38 @@ Example Ansible inventory file
    This is done to prevent secrets from leaking out, for example in ``ps`` output.
 
    We recommend using SSH Keys, and if needed an ssh-agent, rather than passwords, where ever possible.
+
+Miscellaneous Issues
+====================
+
+
+Intermittent failure while using ``network_cli`` connection type
+----------------------------------------------------------------
+
+If the command prompt received in response is not matched correctly within
+the ``network_cli`` connection plugin the task might fail intermittently with truncated
+response or with the error message ``operation requires privilege escalation``.
+Starting in 2.7.1 a new buffer read timer is added to ensure prompts are matched properly
+and a complete response is send in output. The timer default value is 0.2 seconds and
+can be adjusted on a per task basis or can be set globally in seconds.
+
+Example Per task timer setting
+
+.. code-block:: yaml
+
+  - name: gather ios facts
+    ios_facts:
+      gather_subset: all
+    register: result
+    vars:
+      ansible_buffer_read_timeout: 2
+
+
+To make this a global setting, add the following to your ``ansible.cfg`` file:
+
+.. code-block:: ini
+
+   [persistent_connection]
+   buffer_read_timeout = 2
+
+This timer delay per command executed on remote host can be disabled by setting the value to zero.
diff --git a/lib/ansible/plugins/connection/network_cli.py b/lib/ansible/plugins/connection/network_cli.py
index 34151d01b8..6ed540426a 100644
--- a/lib/ansible/plugins/connection/network_cli.py
+++ b/lib/ansible/plugins/connection/network_cli.py
@@ -157,6 +157,21 @@ options:
       - name: ANSIBLE_PERSISTENT_COMMAND_TIMEOUT
     vars:
       - name: ansible_command_timeout
+  persistent_buffer_read_timeout:
+    type: float
+    description:
+      - Configures, in seconds, the amount of time to wait for the data to be read
+        from Paramiko channel after the command prompt is matched. This timeout
+        value ensures that command prompt matched is correct and there is no more data
+        left to be received from remote host.
+    default: 0.1
+    ini:
+      - section: persistent_connection
+        key: buffer_read_timeout
+    env:
+      - name: ANSIBLE_PERSISTENT_BUFFER_READ_TIMEOUT
+    vars:
+      - name: ansible_buffer_read_timeout
 """
 
 import getpass
@@ -164,6 +179,7 @@ import json
 import logging
 import re
 import os
+import signal
 import socket
 import traceback
 
@@ -183,6 +199,10 @@ except ImportError:
     display = Display()
 
 
+class AnsibleCmdRespRecv(Exception):
+    pass
+
+
 class Connection(NetworkConnectionBase):
     ''' CLI (shell) SSH connections on Paramiko '''
 
@@ -199,6 +219,7 @@ class Connection(NetworkConnectionBase):
         self._matched_pattern = None
         self._last_response = None
         self._history = list()
+        self._command_response = None
 
         self._terminal = None
         self.cliconf = None
@@ -342,15 +363,39 @@ class Connection(NetworkConnectionBase):
         '''
         Handles receiving of output from command
         '''
-        recv = BytesIO()
-        handled = False
-
         self._matched_prompt = None
         self._matched_cmd_prompt = None
+        recv = BytesIO()
+        handled = False
+        command_prompt_matched = False
         matched_prompt_window = window_count = 0
 
+        command_timeout = self.get_option('persistent_command_timeout')
+        self._validate_timeout_value(command_timeout, "persistent_command_timeout")
+
+        buffer_read_timeout = self.get_option('persistent_buffer_read_timeout')
+        self._validate_timeout_value(buffer_read_timeout, "persistent_buffer_read_timeout")
+
         while True:
-            data = self._ssh_shell.recv(256)
+            if command_prompt_matched:
+                try:
+                    signal.signal(signal.SIGALRM, self._handle_buffer_read_timeout)
+                    signal.setitimer(signal.ITIMER_REAL, buffer_read_timeout)
+                    data = self._ssh_shell.recv(256)
+                    signal.alarm(0)
+                    # if data is still received on channel it indicates the prompt string
+                    # is wrongly matched in between response chunks, continue to read
+                    # remaining response.
+                    command_prompt_matched = False
+
+                    # restart command_timeout timer
+                    signal.signal(signal.SIGALRM, self._handle_command_timeout)
+                    signal.alarm(command_timeout)
+
+                except AnsibleCmdRespRecv:
+                    return self._command_response
+            else:
+                data = self._ssh_shell.recv(256)
 
             # when a channel stream is closed, received data will be empty
             if not data:
@@ -376,7 +421,11 @@ class Connection(NetworkConnectionBase):
             if self._find_prompt(window):
                 self._last_response = recv.getvalue()
                 resp = self._strip(self._last_response)
-                return self._sanitize(resp, command)
+                self._command_response = self._sanitize(resp, command)
+                if buffer_read_timeout == 0.0:
+                    return self._command_response
+                else:
+                    command_prompt_matched = True
 
     def send(self, command, prompt=None, answer=None, newline=True, sendonly=False, prompt_retry_check=False, check_all=False):
         '''
@@ -398,6 +447,17 @@ class Connection(NetworkConnectionBase):
             display.vvvv(traceback.format_exc(), host=self._play_context.remote_addr)
             raise AnsibleConnectionFailure("timeout trying to send command: %s" % command.strip())
 
+    def _handle_buffer_read_timeout(self, signum, frame):
+        display.vvvv("Response received, triggered 'persistent_buffer_read_timeout' timer of %s seconds"
+                     % self.get_option('persistent_buffer_read_timeout'), host=self._play_context.remote_addr)
+        raise AnsibleCmdRespRecv()
+
+    def _handle_command_timeout(self, signum, frame):
+        msg = 'command timeout triggered, timeout value is %s secs.\nSee the timeout setting options in the Network Debug and Troubleshooting Guide.'\
+              % self.get_option('persistent_command_timeout')
+        display.display(msg, log_only=True)
+        raise AnsibleConnectionFailure(msg)
+
     def _strip(self, data):
         '''
         Removes ANSI codes from device response
@@ -488,3 +548,7 @@ class Connection(NetworkConnectionBase):
             raise AnsibleConnectionFailure(errored_response)
 
         return False
+
+    def _validate_timeout_value(self, timeout, timer_name):
+        if timeout < 0:
+            raise AnsibleConnectionFailure("'%s' timer value '%s' is invalid, value should be greater than or equal to zero." % (timer_name, timeout))
diff --git a/test/units/plugins/connection/test_network_cli.py b/test/units/plugins/connection/test_network_cli.py
index 9831b5d6e6..554b142044 100644
--- a/test/units/plugins/connection/test_network_cli.py
+++ b/test/units/plugins/connection/test_network_cli.py
@@ -21,6 +21,7 @@ from __future__ import (absolute_import, division, print_function)
 __metaclass__ = type
 
 import re
+import time
 import json
 
 from io import StringIO
@@ -28,6 +29,7 @@ from io import StringIO
 from ansible.compat.tests import unittest
 from ansible.compat.tests.mock import patch, MagicMock
 
+from ansible.module_utils._text import to_text
 from ansible.errors import AnsibleConnectionFailure
 from ansible.playbook.play_context import PlayContext
 from ansible.plugins.connection import network_cli
@@ -140,15 +142,14 @@ class TestConnectionClass(unittest.TestCase):
         device#
         """
 
-        mock__shell.recv.return_value = response
-
+        mock__shell.recv.side_effect = [response, None]
         output = conn.send(b'command', None, None, None)
 
         mock__shell.sendall.assert_called_with(b'command\r')
-        self.assertEqual(output, 'command response')
+        self.assertEqual(to_text(conn._command_response), 'command response')
 
         mock__shell.reset_mock()
-        mock__shell.recv.return_value = b"ERROR: error message device#"
+        mock__shell.recv.side_effect = [b"ERROR: error message device#"]
 
         with self.assertRaises(AnsibleConnectionFailure) as exc:
             conn.send(b'command', None, None, None)