Merge "Increase timeout to 20s for Openstack collectd plugins"

deployment_scripts/puppet/modules/lma_collector/README.md

@@ -952,7 +952,10 @@ The resource title should be set to the service name (e.g. `'nova'`).
   a string.
 * `timeout`: *Optional*. Timeout in seconds beyond which the collector
   considers that the endpoint doesn't respond. Valid options: an integer.
-  Default: 5.
+  Default: 20.
+* `max_retries`: *Optional*. Number of maximum retries when an error occurs
+  (including timeout). Valid options: an integer.
+  Default: 2.
 * `pacemaker_master_resource`: *Optional*. Name of the pacemaker resource used
   to determine if the collecting of statistics should be active. This is
   a parameter for advanced users. For this to function the

deployment_scripts/puppet/modules/lma_collector/files/collectd/collectd_openstack.py

@@ -151,6 +151,10 @@ class CollectdPlugin(base.Base):
 
     def __init__(self, *args, **kwargs):
         super(CollectdPlugin, self).__init__(*args, **kwargs)
+        # The timeout/max_retries are defined according to the observations on
+        # 200 nodes environments with 600 VMs. See #1554502 for details.
+        self.timeout = 20
+        self.max_retries = 2
         self.os_client = None
         self.extra_config = {}
 

deployment_scripts/puppet/modules/lma_collector/manifests/collectd/openstack.pp

@@ -18,7 +18,8 @@ define lma_collector::collectd::openstack (
   $password,
   $tenant,
   $keystone_url,
-  $timeout = undef,
+  $timeout = 20,
+  $max_retries = 2,
   $pacemaker_master_resource = undef,
 ) {
 
@@ -32,17 +33,16 @@ define lma_collector::collectd::openstack (
     fail("service '${service}' is not supported")
   }
 
-  $real_timeout = $timeout ? {
-    undef   => $lma_collector::params::openstack_client_timeout,
-    default => $timeout,
-  }
+  validate_integer($timeout)
+  validate_integer($max_retries)
 
   $config = {
     'Username'    => "\"${user}\"",
     'Password'    => "\"${password}\"",
     'Tenant'      => "\"${tenant}\"",
     'KeystoneUrl' => "\"${keystone_url}\"",
-    'Timeout'     => "\"${real_timeout}\"",
+    'Timeout'     => "\"${timeout}\"",
+    'MaxRetries'  => "\"${max_retries}\"",
   }
 
   if $pacemaker_master_resource {

deployment_scripts/puppet/modules/lma_collector/spec/defines/lma_collector_collectd_openstack_spec.rb

@@ -28,18 +28,21 @@ describe 'lma_collector::collectd::openstack' do
              .with_config({"Username" => '"user"', "Password" => '"password"',
                            "Tenant" => '"tenant"',
                            "KeystoneUrl" => '"http://example.com/keystone"',
-                           "Timeout" => '"5"'}) }
+                           "Timeout" => '"20"',
+                           "MaxRetries" => '"2"'}) }
     end
 
     describe 'with required and optional params' do
         let(:title) { :nova }
         let(:params) {{:user => "user", :password => "password", :tenant => "tenant",
                        :keystone_url => "http://example.com/keystone",
-                       :timeout => 10, :pacemaker_master_resource => "vip__management"}}
+                       :timeout => 10, :max_retries => 1,
+                       :pacemaker_master_resource => "vip__management"}}
         it { is_expected.to contain_lma_collector__collectd__python('openstack_nova') \
              .with_config({"Username" => '"user"', "Password" => '"password"',
                            "Tenant" => '"tenant"',
                            "KeystoneUrl" => '"http://example.com/keystone"',
-                           "Timeout" => '"10"', "DependsOnResource" => '"vip__management"'}) }
+                           "Timeout" => '"10"', "MaxRetries" => '"1"',
+                           "DependsOnResource" => '"vip__management"'}) }
     end
 end