Add config parameters for WinRM timeouts

Both the operation and the read timeout can now be configured in the
zuul-executor main configuration file. If the network is flaky,
increasing these numbers from their defaults might help to lower the
rate of aborted Windows builds.

Change-Id: I4c25ca6027fc4150ec1c9c49ed286e7b4f20d4dd

doc/source/admin/components.rst

@@ -522,6 +522,17 @@ The following sections of ``zuul.conf`` are used by the executor:
       .. note:: Currently certificate verification is disabled when
                 connecting to Windows nodes via winrm.
 
+   .. attr:: winrm_operation_timeout_sec
+      :default: None. The Ansible default of 20 is used in this case.
+
+      The timeout for WinRM operations.
+
+   .. attr:: winrm_read_timeout_sec
+      :default: None. The Ansible default of 30 is used in this case.
+
+      The timeout for WinRM read. Increase this if there are intermittent
+      network issues and read timeout errors keep occurring.
+
    .. _admin_sitewide_variables:
 
    .. attr:: variables

tests/fixtures/zuul-winrm.conf

@@ -0,0 +1,33 @@
+[gearman]
+server=127.0.0.1
+
+[statsd]
+# note, use 127.0.0.1 rather than localhost to avoid getting ipv6
+# see: https://github.com/jsocol/pystatsd/issues/61
+server=127.0.0.1
+
+[scheduler]
+tenant_config=main.yaml
+
+[merger]
+git_dir=/tmp/zuul-test/merger-git
+git_user_email=zuul@example.com
+git_user_name=zuul
+
+[executor]
+git_dir=/tmp/zuul-test/executor-git
+winrm_operation_timeout_sec=120
+winrm_read_timeout_sec=180
+
+[connection gerrit]
+driver=gerrit
+server=review.example.com
+user=jenkins
+sshkey=fake_id_rsa_path
+
+[connection smtp]
+driver=smtp
+server=localhost
+port=25
+default_from=zuul@example.com
+default_to=you@example.com

tests/unit/test_inventory.py

@@ -19,12 +19,12 @@ import yaml
 from tests.base import ZuulTestCase
 
 
-class TestInventory(ZuulTestCase):
+class TestInventoryBase(ZuulTestCase):
 
     tenant_config_file = 'config/inventory/main.yaml'
 
     def setUp(self):
-        super(TestInventory, self).setUp()
+        super(TestInventoryBase, self).setUp()
         self.executor_server.hold_jobs_in_build = True
         A = self.fake_gerrit.addFakeChange('org/project', 'master', 'A')
         self.fake_gerrit.addEvent(A.getPatchsetCreatedEvent(1))
@@ -41,6 +41,9 @@ class TestInventory(ZuulTestCase):
                                       'setup-inventory.yaml')
         return yaml.safe_load(open(setup_inv_path, 'r'))
 
+
+class TestInventory(TestInventoryBase):
+
     def test_single_inventory(self):
 
         inventory = self._get_build_inventory('single-inventory')
@@ -157,3 +160,21 @@ class TestInventory(ZuulTestCase):
 
         self.executor_server.release()
         self.waitUntilSettled()
+
+
+class TestWindowsInventory(TestInventoryBase):
+    config_file = 'zuul-winrm.conf'
+
+    def test_windows_inventory(self):
+        inventory = self._get_build_inventory('hostvars-inventory')
+        windows_host = inventory['all']['hosts']['windows']
+        self.assertEqual(windows_host['ansible_connection'], 'winrm')
+        self.assertEqual(
+            windows_host['ansible_winrm_operation_timeout_sec'],
+            '120')
+        self.assertEqual(
+            windows_host['ansible_winrm_read_timeout_sec'],
+            '180')
+
+        self.executor_server.release()
+        self.waitUntilSettled()

zuul/executor/server.py

@@ -661,6 +661,14 @@ class AnsibleJob(object):
         self.winrm_pem_file = get_default(self.executor_server.config,
                                           'executor', 'winrm_cert_pem_file',
                                           '~/.winrm/winrm_client_cert.pem')
+        self.winrm_operation_timeout = get_default(
+            self.executor_server.config,
+            'executor',
+            'winrm_operation_timeout_sec')
+        self.winrm_read_timeout = get_default(
+            self.executor_server.config,
+            'executor',
+            'winrm_read_timeout_sec')
         self.ssh_agent = SshAgent()
 
         self.executor_variables_file = None
@@ -1243,6 +1251,12 @@ class AnsibleJob(object):
                         # now.
                         host_vars['ansible_winrm_server_cert_validation'] = \
                             'ignore'
+                        if self.winrm_operation_timeout is not None:
+                            host_vars['ansible_winrm_operation_timeout_sec'] =\
+                                self.winrm_operation_timeout
+                        if self.winrm_read_timeout is not None:
+                            host_vars['ansible_winrm_read_timeout_sec'] = \
+                                self.winrm_read_timeout
 
                 host_keys = []
                 for key in node.get('host_keys'):