Do not create placeholder status file for AKS extensions (#2298)

azurelinuxagent/ga/exthandlers.py

@@ -720,7 +720,8 @@ def handle_enable(self, ext_handler_i, extension):
         # failures back to CRP. If a placeholder for an extension already exists with Transitioning status, we would
         # not override it, hence we only create a placeholder for enable/disable commands but the extensions have the
         # data to create their own if needed.
-        ext_handler_i.create_placeholder_status_file(extension)
+        if ext_handler_i.should_create_default_placeholder(extension):
+            ext_handler_i.create_placeholder_status_file(extension)
         self.__handle_extension(ext_handler_i, extension, uninstall_exit_code)
 
     @staticmethod
@@ -1389,6 +1390,17 @@ def initialize(self):
         CGroupConfigurator.get_instance().setup_extension_slice(
             extension_name=self.get_full_name())
 
+    def should_create_default_placeholder(self, extension=None):
+        """
+        There's a bug in the AKS extension where they dont update the status file if it exists.
+        This violates the contract we have with extensions. Until they fix their extension,
+        we're going to skip creating a placeholder for them to ensure they dont have any downtime.
+        For all other extensions, we should create a
+        """
+
+        ignore_extension_regex = r"Microsoft.AKS.Compute.AKS\S*"
+        return re.match(ignore_extension_regex, self.get_extension_full_name(extension)) is None
+
     def create_placeholder_status_file(self, extension=None, status=ValidHandlerStatus.transitioning, code=0,
                                        operation="Enabling Extension", message="Install/Enable is in progress."):
         _, status_path = self.get_status_file_path(extension)

tests/data/wire/ext_conf_aks_extension.xml

@@ -0,0 +1,69 @@
+<Extensions version="1.0.0.0" goalStateIncarnation="9">
+    <GuestAgentExtension>
+        <GAFamilies>
+            <GAFamily>
+                <Name>Prod</Name>
+                <Uris>
+                    <Uri>http://mock-goal-state/manifest_of_ga.xml</Uri>
+                </Uris>
+            </GAFamily>
+            <GAFamily>
+                <Name>Test</Name>
+                <Uris>
+                    <Uri>http://mock-goal-state/manifest_of_ga.xml</Uri>
+                </Uris>
+            </GAFamily>
+        </GAFamilies>
+    </GuestAgentExtension>
+    <Plugins>
+        <Plugin name="OSTCExtensions.ExampleHandlerLinux" version="1.0.0" location="http://mock-goal-state/rdfepirv2hknprdstr03.blob.core.windows.net/b01058962be54ceca550a390fa5ff064/Microsoft.OSTCExtensions_ExampleHandlerLinux_asiaeast_manifest.xml" config="" state="enabled" autoUpgrade="false" failoverlocation="http://mock-goal-state/rdfepirv2hknprdstr04.blob.core.windows.net/b01058962be54ceca550a390fa5ff064/Microsoft.OSTCExtensions_ExampleHandlerLinux_asiaeast_manifest.xml" runAsStartupTask="false" isJson="true"/>
+        <Plugin name="Microsoft.AKS.Compute.AKS.Linux.AKSNode" version="1.0.0" location="https://mock-goal-state/zrdfepirv2sn3prdstr06.blob.core.windows.net/3055d769a0ed485d8c5ea8a7cd24073e/Microsoft.Powershell_ExampleExtension_ussouth_manifest.xml" state="enabled" autoUpgrade="true" failoverlocation="https://mock-goal-state/zrdfepirv2sn4prdstr02a.blob.core.windows.net/3055d769a0ed485d8c5ea8a7cd24073e/Microsoft.Powershell_ExampleExtension_ussouth_manifest.xml" runAsStartupTask="false" isJson="true" useExactVersion="false"/>
+        <Plugin name="Microsoft.AKS.Compute.AKS-Engine.Linux.Billing" version="1.0.0" location="https://mock-goal-state/rdfepirv2sn3prdstr01.blob.core.windows.net/8b23413e0ce64aa89889a295b6122112/Microsoft.EnterpriseCloud.Monitoring_ExampleHandlerLinux_ussouth_manifest.xml" state="enabled" autoUpgrade="true" failoverlocation="https://mock-goal-state/ardfepirv2sn3prdstr05.blob.core.windows.net/8b23413e0ce64aa89889a295b6122112/Microsoft.EnterpriseCloud.Monitoring_ExampleHandlerLinux_ussouth_manifest.xml" runAsStartupTask="false" isJson="true" useExactVersion="false"/>
+        <Plugin name="Microsoft.AKS.Compute.AKS.Linux.Billing" version="1.0.0" location="https://mock-goal-state/zrdfepirv2sn4prdstr09a.blob.core.windows.net/f72653efd9e349ed9842c8b99e4c1712/Microsoft.CPlat.Core_ExampleExtensionLinux_ussouth_manifest.xml" state="enabled" autoUpgrade="true" failoverlocation="https://mock-goal-state/ardfepirv2sn3prdstr05.blob.core.windows.net/f72653efd9e349ed9842c8b99e4c1712/Microsoft.CPlat.Core_ExampleExtensionLinux_ussouth_manifest.xml" runAsStartupTask="false" isJson="true" useExactVersion="false"/>
+    </Plugins>
+    <PluginSettings>
+        <Plugin name="OSTCExtensions.ExampleHandlerLinux" version="1.0.0">
+            <RuntimeSettings seqNo="0">
+                {
+                "runtimeSettings": [
+                {
+                "handlerSettings": {
+                "publicSettings": {"message": "Enabling non-AKS"}
+                }
+                }
+                ]
+                }
+            </RuntimeSettings>
+        </Plugin>
+        <Plugin name="Microsoft.AKS.Compute.AKS.Linux.AKSNode" version="1.0.0">
+            <RuntimeSettings seqNo="0">
+                {
+                "runtimeSettings": [
+                {
+                "handlerSettings": {
+                "publicSettings": {"message": "Enabling AKSNode"}
+                }
+                }
+                ]
+                }
+            </RuntimeSettings>
+        </Plugin>
+        <Plugin name="Microsoft.AKS.Compute.AKS-Engine.Linux.Billing" version="1.0.0">
+            <RuntimeSettings seqNo="0">
+                {
+                "runtimeSettings": [
+                {
+                "handlerSettings": {
+                "publicSettings": {"message": "Enabling AKSBilling"}
+                }
+                }
+                ]
+                }
+            </RuntimeSettings>
+        </Plugin>
+    </PluginSettings>
+    <StatusUploadBlob statusBlobType="BlockBlob">
+        https://test.blob.core.windows.net/vhds/test-cs12.test-cs12.test-cs12.status?sr=b&amp;sp=rw&amp;se=9999-01-01&amp;sk=key1&amp;sv=2014-02-14&amp;sig=hfRh7gzUE7sUtYwke78IOlZOrTRCYvkec4hGZ9zZzXo
+    </StatusUploadBlob>
+</Extensions>
+

tests/ga/test_extension.py

@@ -1575,6 +1575,56 @@ def mock_popen(cmd, *args, **kwargs):
                                             expected_msg="Dependent Extension OSTCExtensions.OtherExampleHandlerLinux did not reach a terminal state within the allowed timeout. Last status was {0}".format(
                                                 ValidHandlerStatus.warning))
 
+    def test_it_should_not_create_placeholder_for_aks_extension(self, mock_http_get, mock_crypt_util, *args):
+        original_popen = subprocess.Popen
+
+        def mock_popen(cmd, *_, **kwargs):
+            if 'env' in kwargs:
+                if ExtensionCommandNames.ENABLE not in cmd:
+                    # To force the test extension to not create a status file on Install, changing command
+                    return original_popen(["echo", "not-enable"], *_, **kwargs)
+
+                seq_no = kwargs['env'][ExtCommandEnvVariable.ExtensionSeqNumber]
+                ext_path = kwargs['env'][ExtCommandEnvVariable.ExtensionPath]
+                status_file_name = "{0}.status".format(seq_no)
+                status_file = os.path.join(ext_path, "status", status_file_name)
+                if "AKS" in cmd:
+                    self.assertFalse(os.path.exists(status_file), "Placeholder file should not be created for AKS")
+                else:
+                    self.assertTrue(os.path.exists(status_file), "Placeholder file should be created for all extensions")
+
+            return original_popen(cmd, *_, **kwargs)
+
+        aks_test_mock = DATA_FILE.copy()
+        aks_test_mock["ext_conf"] = "wire/ext_conf_aks_extension.xml"
+
+        exthandlers_handler, protocol = self._create_mock(mockwiredata.WireProtocolData(aks_test_mock),
+                                                          mock_http_get, mock_crypt_util, *args)
+
+        with patch('azurelinuxagent.common.cgroupapi.subprocess.Popen', side_effect=mock_popen):
+            exthandlers_handler.run()
+            exthandlers_handler.report_ext_handlers_status()
+
+            self._assert_handler_status(protocol.report_vm_status, "Ready", 1, "1.0.0",
+                                        expected_handler_name="OSTCExtensions.ExampleHandlerLinux")
+            self._assert_handler_status(protocol.report_vm_status, "Ready", 1, "1.0.0",
+                                        expected_handler_name="Microsoft.AKS.Compute.AKS.Linux.AKSNode")
+            self._assert_handler_status(protocol.report_vm_status, "Ready", 1, "1.0.0",
+                                        expected_handler_name="Microsoft.AKS.Compute.AKS-Engine.Linux.Billing")
+            # Extension without settings
+            self._assert_handler_status(protocol.report_vm_status, "Ready", 0, "1.0.0",
+                                        expected_handler_name="Microsoft.AKS.Compute.AKS.Linux.Billing")
+
+            self._assert_ext_status(protocol.report_vm_status, ValidHandlerStatus.success, 0,
+                                    expected_handler_name="OSTCExtensions.ExampleHandlerLinux",
+                                    expected_msg="Enabling non-AKS")
+            self._assert_ext_status(protocol.report_vm_status, ValidHandlerStatus.success, 0,
+                                    expected_handler_name="Microsoft.AKS.Compute.AKS.Linux.AKSNode",
+                                    expected_msg="Enabling AKSNode")
+            self._assert_ext_status(protocol.report_vm_status, ValidHandlerStatus.success, 0,
+                                    expected_handler_name="Microsoft.AKS.Compute.AKS-Engine.Linux.Billing",
+                                    expected_msg="Enabling AKSBilling")
+
     def test_it_should_include_part_of_status_in_ext_handler_message(self, mock_http_get, mock_crypt_util, *args):
         """
         Testing scenario when the status file is invalid,