Add scheduler-max-attempts flag

Using affinity/anti-affinity policies sometimes we get into a race
condition on the resources that are available on the hypervisors.
This flag allows to increase the number of retries and hence hosts
to schedule on and therefore allowing the successful scheduling
of instances.

This option was taken from the following URL, while future work
with placement is done to help with scheduling with affinity

https://docs.openstack.org/nova/latest/admin/troubleshooting/affinity-policy-violated.html

Signed-off-by: Arif Ali <arif.ali@canonical.com>
Change-Id: I353dbaa38eb0526014888ede27702b428eb66afd

config.yaml

@@ -617,6 +617,17 @@ options:
       .
       Possible Values are positive integers. Any value less than 1 will be
       treated as 1.
+  scheduler-max-attempts:
+    type: int
+    default:
+    description: |
+      The value to be configured for max_attempts property under scheduler.
+      .
+      This is useful for rescheduling instances to hosts when affinity policies
+      are in place as described in the following URL
+      https://docs.openstack.org/nova/latest/admin/troubleshooting/affinity-policy-violated.html
+      .
+      The default 3
   spice-agent-enabled:  # LP: #1856602
     type: boolean
     default: True  # OpenStack's default value.

hooks/nova_cc_context.py

@@ -440,6 +440,8 @@ class NovaConfigContext(ch_context.WorkerConfigContext):
         ctxt['console_access_port'] = hookenv.config('console-access-port')
         ctxt['scheduler_host_subset_size'] = hookenv.config(
             'scheduler-host-subset-size')
+        ctxt['scheduler_max_attempts'] = hookenv.config(
+            'scheduler-max-attempts')
         ctxt['unique_server_names'] = hookenv.config('unique-server-names')
         ctxt['skip_hosts_with_build_failures'] = hookenv.config(
             'skip-hosts-with-build-failures')

templates/ocata/nova.conf

@@ -205,6 +205,10 @@ enabled_filters = {{ scheduler_default_filters }}
 host_subset_size = {{ scheduler_host_subset_size }}
 {%- endif %}
 
+{%- if scheduler_max_attempts %}
+max_attempts = {{ scheduler_max_attempts }}
+{%- endif %}
+
 [api]
 auth_strategy=keystone
 {% if vendor_data or vendor_data_url -%}

templates/pike/nova.conf

@@ -219,6 +219,10 @@ build_failure_weight_multiplier = 0.0
 host_subset_size = {{ scheduler_host_subset_size }}
 {%- endif %}
 
+{%- if scheduler_max_attempts %}
+max_attempts = {{ scheduler_max_attempts }}
+{%- endif %}
+
 [api]
 auth_strategy=keystone
 {% if vendor_data or vendor_data_url -%}

templates/rocky/nova.conf

@@ -221,6 +221,10 @@ build_failure_weight_multiplier = 0.0
 host_subset_size = {{ scheduler_host_subset_size }}
 {%- endif %}
 
+{%- if scheduler_max_attempts %}
+max_attempts = {{ scheduler_max_attempts }}
+{%- endif %}
+
 [api]
 auth_strategy=keystone
 {% if vendor_data or vendor_data_url -%}

templates/train/nova.conf

@@ -235,6 +235,10 @@ build_failure_weight_multiplier = 0.0
 host_subset_size = {{ scheduler_host_subset_size }}
 {%- endif %}
 
+{%- if scheduler_max_attempts %}
+max_attempts = {{ scheduler_max_attempts }}
+{%- endif %}
+
 [api]
 auth_strategy=keystone
 {% if vendor_data or vendor_data_url -%}

unit_tests/test_nova_cc_contexts.py

@@ -341,10 +341,13 @@ class NovaComputeContextTests(CharmTestCase):
         mock_config_ip.side_effect = self.test_config.get
         mock_unit_get.return_value = '127.0.0.1'
         self.test_config.set('scheduler-default-filters', 'TestFilter')
+        self.test_config.set('scheduler-max-attempts', 10)
         self.test_config.set('unique-server-names', 'project')
         ctxt = context.NovaConfigContext()()
         self.assertEqual(ctxt['scheduler_default_filters'],
                          self.config('scheduler-default-filters'))
+        self.assertEqual(ctxt['scheduler_max_attempts'],
+                         self.config('scheduler-max-attempts'))
         self.assertEqual(ctxt['cpu_allocation_ratio'],
                          self.config('cpu-allocation-ratio'))
         self.assertEqual(ctxt['ram_allocation_ratio'],