{AKS} az aks nodepool add/update/upgrade: Add new parameter --node-soak-duration

src/azure-cli/azure/cli/command_modules/acs/_help.py

@@ -1515,6 +1515,9 @@
   - name: --drain-timeout
     type: int
     short-summary: When nodes are drain how many minutes to wait for all pods to be evicted
+  - name: --node-soak-duration
+    type: int
+    short-summary: The amount of time (in minutes) to wait after draining a node and before reimaging it and moving on to next node.
   - name: --enable-encryption-at-host
     type: bool
     short-summary: Enable EncryptionAtHost, default value is false.
@@ -1639,6 +1642,9 @@
   - name: --drain-timeout
     type: int
     short-summary: When nodes are drain how many minutes to wait for all pods to be evicted
+  - name: --node-soak-duration
+    type: int
+    short-summary: The amount of time (in minutes) to wait after draining a node and before reimaging it and moving on to next node.
   - name: --node-taints
     type: string
     short-summary: The node taints for the node pool. You can update the existing node taint of a nodepool or create a new node taint for a nodepool. Pass the empty string `""` to remove all taints.
@@ -1681,6 +1687,9 @@
   - name: --drain-timeout
     type: int
     short-summary: When nodes are drain how long to wait for all pods to be evicted
+  - name: --node-soak-duration
+    type: int
+    short-summary: The amount of time (in minutes) to wait after draining a node and before reimaging it and moving on to next node.
   - name: --snapshot-id
     type: string
     short-summary: The source snapshot id used to upgrade this nodepool.

src/azure-cli/azure/cli/command_modules/acs/_params.py

@@ -573,6 +573,7 @@ def load_arguments(self, _):
         c.argument('node_osdisk_size', type=int)
         c.argument('max_surge', validator=validate_max_surge)
         c.argument('drain_timeout', type=int)
+        c.argument('node_soak_duration', type=int)
         c.argument('mode', get_enum_type(node_mode_types))
         c.argument('scale_down_mode', arg_type=get_enum_type(scale_down_modes))
         c.argument('max_pods', type=int, options_list=['--max-pods', '-m'])
@@ -605,6 +606,7 @@ def load_arguments(self, _):
         c.argument('node_taints', validator=validate_nodepool_taints)
         c.argument('max_surge', validator=validate_max_surge)
         c.argument('drain_timeout', type=int)
+        c.argument('node_soak_duration', type=int)
         c.argument('mode', get_enum_type(node_mode_types))
         c.argument('scale_down_mode', arg_type=get_enum_type(scale_down_modes))
         c.argument('allowed_host_ports', nargs='+', validator=validate_allowed_host_ports)
@@ -613,6 +615,7 @@ def load_arguments(self, _):
     with self.argument_context('aks nodepool upgrade') as c:
         c.argument('max_surge', validator=validate_max_surge)
         c.argument('drain_timeout', type=int)
+        c.argument('node_soak_duration', type=int)
         c.argument('snapshot_id', validator=validate_snapshot_id)
         c.argument('yes', options_list=['--yes', '-y'], help='Do not prompt for confirmation.', action='store_true')
 

src/azure-cli/azure/cli/command_modules/acs/agentpool_decorator.py

@@ -972,6 +972,26 @@ def get_drain_timeout(self):
         # this parameter does not need validation
         return drain_timeout
 
+    def get_node_soak_duration(self):
+        """Obtain the value of node_soak_duration.
+
+        :return: int
+        """
+        # read the original value passed by the command
+        node_soak_duration = self.raw_param.get("node_soak_duration")
+        # In create mode, try to read the property value corresponding to the parameter from the `agentpool` object
+        if self.decorator_mode == DecoratorMode.CREATE:
+            if (
+                self.agentpool and
+                self.agentpool.upgrade_settings and
+                self.agentpool.upgrade_settings.node_soak_duration_in_minutes is not None
+            ):
+                node_soak_duration = self.agentpool.upgrade_settings.node_soak_duration_in_minutes
+
+        # this parameter does not need dynamic completion
+        # this parameter does not need validation
+        return node_soak_duration
+
     def get_vm_set_type(self) -> str:
         """Obtain the value of vm_set_type, default value is CONST_VIRTUAL_MACHINE_SCALE_SETS.
 
@@ -1562,6 +1582,10 @@ def set_up_upgrade_settings(self, agentpool: AgentPool) -> AgentPool:
         if drain_timeout:
             upgrade_settings.drain_timeout_in_minutes = drain_timeout
 
+        node_soak_duration = self.context.get_node_soak_duration()
+        if node_soak_duration:
+            upgrade_settings.node_soak_duration_in_minutes = node_soak_duration
+
         agentpool.upgrade_settings = upgrade_settings
         return agentpool
 
@@ -1859,6 +1883,11 @@ def update_upgrade_settings(self, agentpool: AgentPool) -> AgentPool:
             upgrade_settings.drain_timeout_in_minutes = drain_timeout
             agentpool.upgrade_settings = upgrade_settings
 
+        node_soak_duration = self.context.get_node_soak_duration()
+        if node_soak_duration:
+            upgrade_settings.node_soak_duration_in_minutes = node_soak_duration
+            agentpool.upgrade_settings = upgrade_settings
+
         return agentpool
 
     def update_vm_properties(self, agentpool: AgentPool) -> AgentPool:

src/azure-cli/azure/cli/command_modules/acs/custom.py

@@ -2197,6 +2197,7 @@ def aks_agentpool_add(
     node_osdisk_size=None,
     max_surge=None,
     drain_timeout=None,
+    node_soak_duration=None,
     mode=CONST_NODEPOOL_MODE_USER,
     scale_down_mode=CONST_SCALE_DOWN_MODE_DELETE,
     max_pods=None,
@@ -2255,6 +2256,7 @@ def aks_agentpool_update(
     node_taints=None,
     max_surge=None,
     drain_timeout=None,
+    node_soak_duration=None,
     mode=None,
     scale_down_mode=None,
     no_wait=False,
@@ -2295,6 +2297,7 @@ def aks_agentpool_upgrade(cmd, client, resource_group_name, cluster_name,
                           node_image_only=False,
                           max_surge=None,
                           drain_timeout=None,
+                          node_soak_duration=None,
                           snapshot_id=None,
                           no_wait=False,
                           aks_custom_headers=None,
@@ -2312,11 +2315,11 @@ def aks_agentpool_upgrade(cmd, client, resource_group_name, cluster_name,
         )
 
     # Note: we exclude this option because node image upgrade can't accept nodepool put fields like max surge
-    if (max_surge or drain_timeout) and node_image_only:
+    if (max_surge or drain_timeout or node_soak_duration) and node_image_only:
         raise MutuallyExclusiveArgumentError(
-            'Conflicting flags. Unable to specify max-surge/drain-timeout with node-image-only.'
-            'If you want to use max-surge/drain-timeout with a node image upgrade, please first '
-            'update max-surge/drain-timeout using "az aks nodepool update --max-surge/--drain-timeout".'
+            'Conflicting flags. Unable to specify max-surge/drain-timeout/node-soak-duration with node-image-only.'
+            'If you want to use max-surge/drain-timeout/node-soak-duration with a node image upgrade, please first '
+            'update max-surge/drain-timeout/node-soak-duration using "az aks nodepool update --max-surge/--drain-timeout/--node-soak-duration".'
         )
 
     if node_image_only:
@@ -2365,6 +2368,8 @@ def aks_agentpool_upgrade(cmd, client, resource_group_name, cluster_name,
         instance.upgrade_settings.max_surge = max_surge
     if drain_timeout:
         instance.upgrade_settings.drain_timeout_in_minutes = drain_timeout
+    if node_soak_duration:
+        instance.upgrade_settings.node_soak_duration_in_minutes = node_soak_duration
 
     # custom headers
     aks_custom_headers = extract_comma_separated_string(

src/azure-cli/azure/cli/command_modules/acs/tests/latest/test_aks_commands.py

@@ -5890,6 +5890,45 @@ def test_aks_nodepool_drain_timeout(self, resource_group, resource_group_locatio
         self.cmd(
             'aks delete -g {resource_group} -n {name} --yes --no-wait', checks=[self.is_empty()])
 
+    @AllowLargeResponse()
+    @AKSCustomResourceGroupPreparer(random_name_length=17, name_prefix='clitest', location='westus2')
+    def test_aks_nodepool_node_soak_duration(self, resource_group, resource_group_location):
+        aks_name = self.create_random_name('cliakstest', 16)
+        np_name = self.create_random_name('clinp', 12)
+        self.kwargs.update({
+            'name': aks_name,
+            'resource_group': resource_group,
+            'nodepool_name': np_name,
+            'ssh_key_value': self.generate_ssh_keys()
+        })
+
+        # create
+        create_cmd = 'aks create --resource-group={resource_group} --name={name} ' \
+                     '--ssh-key-value={ssh_key_value} -c 1'
+        self.cmd(create_cmd, checks=[
+            self.check('provisioningState', 'Succeeded'),
+        ])
+
+        add_nodepool_cmd = 'aks nodepool add -g {resource_group} --cluster-name {name} -n {nodepool_name} ' \
+                     '--mode user --node-soak-duration 5'
+        self.cmd(add_nodepool_cmd, checks=[
+            self.check('provisioningState', 'Succeeded'),
+            self.check('upgradeSettings.nodeSoakDurationInMinutes', 5),
+        ])
+
+        update_nodepool_cmd = 'aks nodepool update -g {resource_group} --cluster-name {name} -n {nodepool_name} ' \
+                      '--node-soak-duration 10'
+        self.cmd(update_nodepool_cmd, checks=[
+            self.check('provisioningState', 'Succeeded'),
+            self.check('upgradeSettings.nodeSoakDurationInMinutes', 10),
+        ])
+
+        #actually running an upgrade is too expensive for these tests.
+
+        # delete
+        self.cmd(
+            'aks delete -g {resource_group} -n {name} --yes --no-wait', checks=[self.is_empty()])
+
     @AllowLargeResponse()
     @AKSCustomResourceGroupPreparer(random_name_length=17, name_prefix='clitest', location='westcentralus')
     def test_aks_nodepool_stop_and_start(self, resource_group, resource_group_location):

src/azure-cli/requirements.py3.Darwin.txt

@@ -32,7 +32,7 @@ azure-mgmt-cognitiveservices==13.5.0
 azure-mgmt-compute==30.4.0
 azure-mgmt-containerinstance==10.1.0
 azure-mgmt-containerregistry==10.1.0
-azure-mgmt-containerservice==28.0.0
+azure-mgmt-containerservice==29.0.0
 azure-mgmt-core==1.3.2
 azure-mgmt-cosmosdb==9.4.0
 azure-mgmt-databoxedge==1.0.0