add support for multiple partner failover group commands; add automated test cases

Author: alecbain

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

@@ -58,6 +58,7 @@
     DatabaseCapabilitiesAdditionalDetails,
     ElasticPoolCapabilitiesAdditionalDetails,
     FailoverPolicyType,
+    FailoverReadOnlyEndpointPolicy,
     ResourceIdType,
     ServicePrincipalType,
     SqlServerMinimalTlsVersionType,
@@ -1727,6 +1728,10 @@ def _configure_security_policy_storage_params(arg_ctx):
                    arg_type=try_planned_before_forced_failover_param_type)
         c.argument('secondary_type', help="Databases secondary type on partner server",
                    arg_type=get_enum_type(FailoverGroupDatabasesSecondaryType))
+        c.argument('partner_server_ids', nargs='+', help="The list of partner server resource id's of the Failover Group")
+        c.argument('readonly_failover_policy', help="The policy of the read only endpoint of the Failover Group",
+                   arg_type=get_enum_type(FailoverReadOnlyEndpointPolicy))
+        c.argument('readonly_endpoint_target', help="The resource id of the read only endpoint target server")
 
     ###############################################
     #             sql instance pool               #

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

@@ -576,7 +576,7 @@ def _get_database_identity(
 def _failover_group_update_common(
         instance,
         failover_policy=None,
-        grace_period=None,):
+        grace_period=None):
     '''
     Updates the failover group grace period and failover policy. Common logic for both Sterling and Managed Instance
     '''
@@ -644,6 +644,9 @@ class FailoverPolicyType(Enum):
     automatic = 'Automatic'
     manual = 'Manual'
 
+class FailoverReadOnlyEndpointPolicy(Enum):
+    enabled = 'Enabled'
+    disabled = 'Disabled'
 
 class FailoverGroupDatabasesSecondaryType(Enum):
     geo = 'Geo'
@@ -6519,7 +6522,10 @@ def failover_group_create(
         partner_resource_group=None,
         failover_policy=FailoverPolicyType.manual.value,
         grace_period=1,
-        add_db=None):
+        add_db=None,
+        partner_server_ids=None,
+        readonly_failover_policy=FailoverReadOnlyEndpointPolicy.disabled.value,
+        readonly_endpoint_target=None):
     '''
     Creates a failover group.
     '''
@@ -6552,14 +6558,24 @@ def failover_group_create(
         add_db,
         [])
 
+    if partner_server_ids is not None:
+        print(partner_server_ids)
+        partner_servers = [PartnerInfo(id=p) for p in partner_server_ids]
+    else:
+        partner_servers = [partner_server]
+
+    if readonly_endpoint_target is None:
+        readonly_endpoint_target = partner_server_id
+
     failover_group_params = FailoverGroup(
-        partner_servers=[partner_server],
+        partner_servers=partner_servers,
         databases=databases,
         read_write_endpoint=FailoverGroupReadWriteEndpoint(
             failover_policy=failover_policy,
             failover_with_data_loss_grace_period_minutes=grace_period),
         read_only_endpoint=FailoverGroupReadOnlyEndpoint(
-            failover_policy="Disabled")
+            failover_policy=readonly_failover_policy,
+            target_server=readonly_endpoint_target)
     )
 
     if secondary_type is not None:
@@ -6581,7 +6597,10 @@ def failover_group_update(
         failover_policy=None,
         grace_period=None,
         add_db=None,
-        remove_db=None):
+        remove_db=None,
+        readonly_endpoint_target=None,
+        readonly_failover_policy=None,
+        partner_server_ids=None):
     '''
     Updates the failover group.
     '''
@@ -6609,6 +6628,13 @@ def failover_group_update(
     if secondary_type is not None:
         instance.secondary_type = secondary_type
 
+    if partner_server_ids is not None:
+        instance.partner_servers = [PartnerInfo(id=p) for p in partner_server_ids]
+
+    instance.read_only_endpoint = FailoverGroupReadOnlyEndpoint(
+        failover_policy=readonly_failover_policy if readonly_failover_policy is not None else instance.read_only_endpoint.failover_policy,
+        target_server=readonly_endpoint_target if readonly_endpoint_target is not None else instance.readonly_endpoint_target)
+
     return instance
 
 

src/azure-cli/azure/cli/command_modules/sql/tests/latest/test_sql_commands.py

@@ -6804,6 +6804,254 @@ def __init__(self, name, group, location):
                      JMESPathCheck('length(@)', 0)
                  ])
 
+    # create 2 servers in the same resource group, and 1 server in a different resource group
+    @ResourceGroupPreparer(parameter_name="resource_group_1",
+                           parameter_name_for_location="resource_group_location_1")
+    @ResourceGroupPreparer(parameter_name="resource_group_2",
+                           parameter_name_for_location="resource_group_location_2")
+    @SqlServerPreparer(parameter_name="server_name_1",
+                       resource_group_parameter_name="resource_group_1",
+                       location='northeurope')
+    @SqlServerPreparer(parameter_name="server_name_2",
+                       resource_group_parameter_name="resource_group_1", 
+                       location='uksouth')
+    @SqlServerPreparer(parameter_name="server_name_3",
+                       resource_group_parameter_name="resource_group_2", 
+                       location='swedencentral')
+    def test_sql_failover_group_mgmt_multiple_partners(self,
+                                     resource_group_1, resource_group_location_1,
+                                     resource_group_2, resource_group_location_2,
+                                     server_name_1, server_name_2, server_name_3):
+        # helper class so that it's clear which servers are in which groups
+        class ServerInfo:  # pylint disable=too-few-public-methods
+            def __init__(self, name, group, location):
+                self.name = name
+                self.group = group
+                self.location = location
+
+        from azure.cli.core.commands.client_factory import get_subscription_id
+
+        s1 = ServerInfo(server_name_1, resource_group_1, resource_group_location_1)
+        s2 = ServerInfo(server_name_2, resource_group_1, resource_group_location_1)
+        s3 = ServerInfo(server_name_3, resource_group_2, resource_group_location_2)
+
+        failover_group_name = "fgclitest-multiplepartner123789"
+
+        database_name = "db1"
+
+        server2_id = "/subscriptions/{}/resourceGroups/{}/providers/Microsoft.Sql/servers/{}".format(
+            get_subscription_id(self.cli_ctx),
+            resource_group_1,
+            server_name_2)
+
+        server3_id = "/subscriptions/{}/resourceGroups/{}/providers/Microsoft.Sql/servers/{}".format(
+            get_subscription_id(self.cli_ctx),
+            resource_group_2,
+            server_name_3)
+
+        # Create database on primary server
+        self.cmd('sql db create -g {} --server {} --name {}'
+                 .format(s1.group, s1.name, database_name),
+                 checks=[
+                     JMESPathCheck('resourceGroup', s1.group),
+                     JMESPathCheck('name', database_name)
+                 ])
+
+        # Create Failover Group
+        self.cmd(
+            'sql failover-group create -n {} -g {} -s {} --partner-resource-group {} --partner-server {} --failover-policy Automatic --grace-period 2 --partner-server-ids {} {} --readonly-failover-policy Enabled'
+                .format(failover_group_name, s1.group, s1.name, s2.group, s2.name, server2_id, server3_id),
+            checks=[
+                JMESPathCheck('name', failover_group_name),
+                JMESPathCheck('resourceGroup', s1.group),
+                JMESPathCheck('partnerServers[0].id', server2_id),
+                JMESPathCheck('partnerServers[1].id', server3_id),
+                JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 120),
+                JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Enabled'),
+                JMESPathCheck('readOnlyEndpoint.targetServer', server2_id),
+                JMESPathCheck('length(databases)', 0)
+            ])
+
+        # List of all failover groups on the primary server
+        self.cmd('sql failover-group list -g {} -s {}'
+                 .format(s1.group, s1.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 1),
+                     JMESPathCheck('[0].name', failover_group_name),
+                     JMESPathCheck('[0].replicationRole', 'Primary')
+                 ])
+
+        # Get Failover Group on a partner server and check if role is secondary
+        self.cmd('sql failover-group show -g {} -s {} -n {}'
+                 .format(s2.group, s2.name, failover_group_name),
+                 checks=[
+                     JMESPathCheck('name', failover_group_name),
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                     JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 120),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Enabled'),
+                     JMESPathCheck('replicationRole', 'Secondary'),
+                     JMESPathCheck('length(databases)', 0)
+                 ])
+
+        self.cmd('sql failover-group show -g {} -s {} -n {}'
+                .format(s3.group, s3.name, failover_group_name),
+                checks=[
+                    JMESPathCheck('name', failover_group_name),
+                    JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                    JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 120),
+                    JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Enabled'),
+                    JMESPathCheck('replicationRole', 'Secondary'),
+                    JMESPathCheck('length(databases)', 0)
+                ])
+
+        if self.in_recording:
+            time.sleep(60)
+
+        # Update Failover Group
+        self.cmd('sql failover-group update -g {} -s {} -n {} --grace-period 3 --add-db {} --readonly-endpoint-target {} --readonly-failover-policy Disabled'
+                 .format(s1.group, s1.name, failover_group_name, database_name, server3_id),
+                 checks=[
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                     JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 180),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('readOnlyEndpoint.targetServer', server3_id),
+                     JMESPathCheck('length(databases)', 1)
+                 ])
+
+        # Check if properties got propagated to secondary server
+        self.cmd('sql failover-group show -g {} -s {} -n {}'
+                 .format(s2.group, s2.name, failover_group_name),
+                 checks=[
+                     JMESPathCheck('name', failover_group_name),
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                     JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 180),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('readOnlyEndpoint.targetServer', server3_id),
+                     JMESPathCheck('replicationRole', 'Secondary'),
+                     JMESPathCheck('length(databases)', 1)
+                 ])
+
+        self.cmd('sql failover-group show -g {} -s {} -n {}'
+                 .format(s3.group, s3.name, failover_group_name),
+                 checks=[
+                     JMESPathCheck('name', failover_group_name),
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                     JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 180),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('readOnlyEndpoint.targetServer', server3_id),
+                     JMESPathCheck('replicationRole', 'Secondary'),
+                     JMESPathCheck('length(databases)', 1)
+                 ])
+
+        # Check if database is created on partner side
+        self.cmd('sql db list -g {} -s {}'
+                 .format(s2.group, s2.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 2)
+                 ])
+
+        self.cmd('sql db list -g {} -s {}'
+                 .format(s3.group, s3.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 2)
+                 ])
+
+        if self.in_recording:
+            time.sleep(60)
+
+        # Check secondary type parameter functionality
+        self._test_failover_group_with_secondary_type(s1, s2, failover_group_name, "Standby")
+
+        self._test_failover_group_with_secondary_type(s1, s2, failover_group_name, "Geo")
+
+        # Update Failover Group failover policy to Manual
+        self.cmd('sql failover-group update -g {} -s {} -n {} --failover-policy Manual'
+                 .format(s1.group, s1.name, failover_group_name),
+                 checks=[
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Manual'),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('length(databases)', 1)
+                 ])
+
+        # Failover failover group from secondary server and then fail back
+        self._test_failover_group_failover(s1, s2, failover_group_name, self.FailoverType.planned)
+
+        self._test_failover_group_failover(s1, s2, failover_group_name, self.FailoverType.forced)
+
+        self._test_failover_group_failover(s1, s2, failover_group_name, self.FailoverType.hybrid)
+
+        self._test_failover_group_failover(s1, s3, failover_group_name, self.FailoverType.planned)
+
+        self._test_failover_group_failover(s1, s3, failover_group_name, self.FailoverType.forced)
+
+        self._test_failover_group_failover(s1, s3, failover_group_name, self.FailoverType.hybrid)
+
+        # Failover failover group from primary server (No-op)
+        self._test_failover_group_failover_from_primary(s1, s2, failover_group_name, self.FailoverType.planned)
+
+        self._test_failover_group_failover_from_primary(s1, s2, failover_group_name, self.FailoverType.forced)
+
+        self._test_failover_group_failover_from_primary(s1, s2, failover_group_name, self.FailoverType.hybrid)
+
+        # Remove database from failover group
+        self.cmd('sql failover-group update -g {} -s {} -n {} --remove-db {}'
+                 .format(s1.group, s1.name, failover_group_name, database_name),
+                 checks=[
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Manual'),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('length(databases)', 0)
+                 ])
+
+        # Check if database got removed
+        self.cmd('sql db show -g {} -s {} -n {}'
+                 .format(s2.group, s2.name, database_name),
+                 checks=[
+                     JMESPathCheck('[0].failoverGroupId', 'None')
+                 ])
+
+        self.cmd('sql db show -g {} -s {} -n {}'
+                 .format(s3.group, s3.name, database_name),
+                 checks=[
+                     JMESPathCheck('[0].failoverGroupId', 'None')
+                 ])
+
+        # Remove partner server from failover group
+        self.cmd('sql failover-group update -g {} -s {} -n {} --partner-server-ids {}'
+                 .format(s1.group, s1.name, failover_group_name, server3_id),
+                 checks=[
+                     JMESPathCheck('partnerServers[0].id', server3_id),
+                     JMESPathCheck('readWriteEndpoint.failoverPolicy', 'Automatic'),
+                     JMESPathCheck('readWriteEndpoint.failoverWithDataLossGracePeriodMinutes', 180),
+                     JMESPathCheck('readOnlyEndpoint.failoverPolicy', 'Disabled'),
+                     JMESPathCheck('readOnlyEndpoint.targetServer', server3_id),
+                     JMESPathCheck('length(databases)', 1)
+                 ])
+
+        # Check partner server was removed
+        self.cmd('sql failover-group list -g {} -s {}'
+                 .format(s2.group, s2.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 0)
+                 ])
+
+        # Drop failover group
+        self.cmd('sql failover-group delete -g {} -s {} -n {}'
+                 .format(s1.group, s1.name, failover_group_name))
+
+        # Check if failover group really got dropped
+        self.cmd('sql failover-group list -g {} -s {}'
+                 .format(s1.group, s1.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 0)
+                 ])
+
+        self.cmd('sql failover-group list -g {} -s {}'
+                 .format(s3.group, s3.name),
+                 checks=[
+                     JMESPathCheck('length(@)', 0)
+                 ])
+
 class SqlVirtualClusterMgmtScenarioTest(ScenarioTest):
     @ManagedInstancePreparer()
     def test_sql_virtual_cluster_mgmt(self, mi, rg):