Merge pull request #652 from kubernetes-incubator/debug_mode

Tune dnsmasq/kubedns limits, replicas, logging

docs/large-deployments.md

@@ -21,5 +21,10 @@ For a large scaled deployments, consider the following configuration changes:
   load on a delegate (the first K8s master node) then retrying failed
   push or download operations.
 
+* Tune parameters for DNS related applications (dnsmasq daemon set, kubedns
+  replication controller). Those are ``dns_replicas``, ``dns_cpu_limit``,
+  ``dns_cpu_requests``, ``dns_memory_limit``, ``dns_memory_requests``.
+  Please note that limits must always be greater than or equal to requests.
+
 For example, when deploying 200 nodes, you may want to run ansible with
 ``--forks=50``, ``--timeout=600`` and define the ``retry_stagger: 60``.

roles/dnsmasq/defaults/main.yml

@@ -27,3 +27,9 @@ skip_dnsmasq: false
 
 # Skip setting up dnsmasq daemonset
 skip_dnsmasq_k8s: "{{ skip_dnsmasq }}"
+
+# Limits for dnsmasq/kubedns apps
+dns_cpu_limit: 100m
+dns_memory_limit: 170Mi
+dns_cpu_requests: 70m
+dns_memory_requests: 70Mi

roles/dnsmasq/templates/01-kube-dns.conf.j2

@@ -20,6 +20,9 @@ server=169.254.169.254
  server=8.8.4.4
 {% endif %}
 
+{% if kube_log_level == 4 %}
+log-queries
+{% endif %}
 bogus-priv
 no-resolv
 no-negcache

roles/dnsmasq/templates/dnsmasq-ds.yml

@@ -29,8 +29,11 @@ spec:
           imagePullPolicy: IfNotPresent
           resources:
             limits:
-              cpu: 100m
+              cpu: {{ dns_cpu_limit }}
-              memory: 256M
+              memory: {{ dns_memory_limit }}
+            requests:
+              cpu: {{ dns_cpu_requests }}
+              memory: {{ dns_memory_requests }}
           ports:
             - name: dns
               containerPort: 53

roles/kubernetes-apps/ansible/defaults/main.yml

@@ -3,6 +3,13 @@ kubedns_version: 1.7
 kubednsmasq_version: 1.3
 exechealthz_version: 1.1
 
+# Limits for dnsmasq/kubedns apps
+dns_cpu_limit: 100m
+dns_memory_limit: 170Mi
+dns_cpu_requests: 70m
+dns_memory_requests: 70Mi
+dns_replicas: 1
+
 # Images
 kubedns_image_repo: "gcr.io/google_containers/kubedns-amd64"
 kubedns_image_tag: "{{ kubedns_version }}"

roles/kubernetes-apps/ansible/templates/kubedns-rc.yml

@@ -8,7 +8,7 @@ metadata:
     version: v19
     kubernetes.io/cluster-service: "true"
 spec:
-  replicas: 1
+  replicas: {{ dns_replicas }}
   selector:
     k8s-app: kubedns
     version: v19
@@ -29,11 +29,11 @@ spec:
           # guaranteed class. Currently, this container falls into the
           # "burstable" category so the kubelet doesn't backoff from restarting it.
           limits:
-            cpu: 100m
+            cpu: {{ dns_cpu_limit }}
-            memory: 170Mi
+            memory: {{ dns_memory_limit }}
           requests:
-            cpu: 100m
+            cpu: {{ dns_cpu_requests }}
-            memory: 70Mi
+            memory: {{ dns_memory_requests }}
         livenessProbe:
           httpGet:
             path: /healthz
@@ -56,6 +56,7 @@ spec:
         # command = "/kube-dns"
         - --domain={{ dns_domain }}.
         - --dns-port=10053
+        - --v={{ kube_log_level }}
         ports:
         - containerPort: 10053
           name: dns-local
@@ -66,11 +67,21 @@ spec:
       - name: dnsmasq
         image: "{{ kubednsmasq_image_repo }}:{{ kubednsmasq_image_tag }}"
         imagePullPolicy: {{ k8s_image_pull_policy }}
+        resources:
+          limits:
+            cpu: {{ dns_cpu_limit }}
+            memory: {{ dns_memory_limit }}
+          requests:
+            cpu: {{ dns_cpu_requests }}
+            memory: {{ dns_memory_requests }}
         args:
         - --log-facility=-
         - --cache-size=1000
         - --no-resolv
         - --server=127.0.0.1#10053
+{% if kube_log_level == 4 %}
+        - --log-queries
+{% endif %}
         ports:
         - containerPort: 53
           name: dns

roles/kubernetes/master/templates/manifests/kube-apiserver.manifest.j2

@@ -41,7 +41,7 @@ spec:
 {% if enable_network_policy is defined and enable_network_policy == True %}
     - --runtime-config=extensions/v1beta1/networkpolicies=true
 {% endif %}
-    - --v={{ kube_log_level | default('2') }}
+    - --v={{ kube_log_level }}
     - --allow-privileged=true
 {% if cloud_provider is defined and cloud_provider == "openstack" %}
     - --cloud-provider={{ cloud_provider }}

roles/kubernetes/master/templates/manifests/kube-controller-manager.manifest.j2

@@ -19,7 +19,7 @@ spec:
     - --service-account-private-key-file={{ kube_cert_dir }}/apiserver-key.pem
     - --root-ca-file={{ kube_cert_dir }}/ca.pem
     - --enable-hostpath-provisioner={{ kube_hostpath_dynamic_provisioner }}
-    - --v={{ kube_log_level | default('2') }}
+    - --v={{ kube_log_level }}
 {% if cloud_provider is defined and cloud_provider == "openstack" %}
     - --cloud-provider={{cloud_provider}}
     - --cloud-config={{ kube_config_dir }}/cloud_config

roles/kubernetes/master/templates/manifests/kube-scheduler.manifest.j2

@@ -16,7 +16,7 @@ spec:
     - scheduler
     - --leader-elect=true
     - --master={{ kube_apiserver_endpoint }}
-    - --v={{ kube_log_level | default('2') }}
+    - --v={{ kube_log_level }}
     livenessProbe:
       httpGet:
         host: 127.0.0.1

roles/kubernetes/node/templates/kubelet.j2

@@ -5,7 +5,7 @@ KUBE_LOGGING="--log-dir={{ kube_log_dir }} --logtostderr=true"
 # logging to stderr means we get it in the systemd journal
 KUBE_LOGGING="--logtostderr=true"
 {% endif %}
-KUBE_LOG_LEVEL="--v={{ kube_log_level | default('2') }}"
+KUBE_LOG_LEVEL="--v={{ kube_log_level }}"
 # The address for the info server to serve on (set to 0.0.0.0 or "" for all interfaces)
 KUBELET_ADDRESS="--address={{ ip | default("0.0.0.0") }}"
 # The port for the info server to serve on

roles/kubernetes/node/templates/manifests/kube-proxy.manifest.j2

@@ -14,7 +14,7 @@ spec:
     command:
     - /hyperkube
     - proxy
-    - --v={{ kube_log_level | default('2') }}
+    - --v={{ kube_log_level }}
     - --master={{ kube_apiserver_endpoint }}
 {% if not is_kube_master %}
     - --kubeconfig=/etc/kubernetes/node-kubeconfig.yaml