Move nodeSelector, affinity and tolerations from templates to values (#45)

* Move nodeSelector, affinity and tolerations from templates to values

* Fix typo

* Fix readme and default values
Add default datanode affinity
Add emptyline to the end of journalnode values file
Fix namenode readme

Author: vvbogdanov87

charts/hdfs-datanode-k8s/templates/datanode-daemonset.yaml

@@ -25,6 +25,18 @@ spec:
       labels:
         name: hdfs-datanode
     spec:
+      {{- if .Values.affinity }}
+      affinity:
+{{ toYaml .Values.affinity | indent 8 }}
+      {{- end }}
+      {{- if .Values.nodeSelector }}
+      nodeSelector:
+{{ toYaml .Values.nodeSelector | indent 8 }}
+      {{- end }}
+      {{- if .Values.tolerations }}
+      tolerations:
+{{ toYaml .Values.tolerations | indent 8 }}
+      {{- end }}
       hostNetwork: true
       hostPID: true
       dnsPolicy: ClusterFirstWithHostNet
@@ -164,13 +176,6 @@ spec:
             - name: jsvc-home
               mountPath: /jsvc-home
       {{- end }}
-      affinity:
-        nodeAffinity:
-          requiredDuringSchedulingIgnoredDuringExecution:
-            nodeSelectorTerms:
-              - matchExpressions:
-                - key: hdfs-datanode-exclude
-                  operator: DoesNotExist
       restartPolicy: Always
       volumes:
         {{- range $index, $path := .Values.dataNodeHostPath }}

charts/hdfs-datanode-k8s/values.yaml

@@ -62,3 +62,14 @@ jsvcEnabled: true
 
 # Whether or not to expect namenodes in the HA setup.
 namenodeHAEnabled: true
+
+# Node labels and tolerations for pod assignment
+nodeSelector: {}
+tolerations: []
+affinity:
+  nodeAffinity:
+    requiredDuringSchedulingIgnoredDuringExecution:
+      nodeSelectorTerms:
+        - matchExpressions:
+          - key: hdfs-datanode-exclude
+            operator: DoesNotExist

charts/hdfs-journalnode-k8s/templates/journalnode-statefulset.yaml

@@ -60,16 +60,18 @@ spec:
       labels:
         app: hdfs-journalnode
     spec:
+      {{- if .Values.affinity }}
       affinity:
-        podAntiAffinity:
-          requiredDuringSchedulingIgnoredDuringExecution:
-            - labelSelector:
-                matchExpressions:
-                  - key: "app"
-                    operator: In
-                    values:
-                    - hfds-journalnode
-              topologyKey: "kubernetes.io/hostname"
+{{ toYaml .Values.affinity | indent 8 }}
+      {{- end }}
+      {{- if .Values.nodeSelector }}
+      nodeSelector:
+{{ toYaml .Values.nodeSelector | indent 8 }}
+      {{- end }}
+      {{- if .Values.tolerations }}
+      tolerations:
+{{ toYaml .Values.tolerations | indent 8 }}
+      {{- end }}
       containers:
         - name: hdfs-journalnode
           image: uhopper/hadoop-namenode:2.7.2

charts/hdfs-journalnode-k8s/values.yaml

@@ -50,3 +50,17 @@ kerberosConfigFileName: krb5.conf
 # where HOST-NAME should match the cluster node
 # host name that each per-host hdfs principal is associated with.
 kerberosKeytabsSecret: hdfs-kerberos-keytabs
+
+# Node labels and tolerations for pod assignment
+nodeSelector: {}
+tolerations: []
+affinity:
+  podAntiAffinity:
+    requiredDuringSchedulingIgnoredDuringExecution:
+      - labelSelector:
+          matchExpressions:
+            - key: "app"
+              operator: In
+              values:
+              - hfds-journalnode
+        topologyKey: "kubernetes.io/hostname"

charts/hdfs-namenode-k8s/README.md

@@ -132,12 +132,12 @@ See the other chart for `datanodes`.
      The two variables above are required. For other variables, see values.yaml.
 
      If also using namenode labels for Kerberos, add
-     the namenodePinningEnabled option:
+     the nodeSelector option:
      ```
      $ helm install -n my-hdfs-namenode  \
          --set kerberosEnabled=true  \
          --set kerberosRealm=MYCOMPANY.COM  \
-         --set namenodePinningEnabled=true \
+         --set nodeSelector.hdfs-namenode-selector=hdfs-namenode \
          hdfs-namenode-k8s
      ```
 

charts/hdfs-namenode-k8s/templates/namenode-statefulset.yaml

@@ -117,16 +117,18 @@ spec:
       {{- else }}
       dnsPolicy: ClusterFirst
       {{- end }}
+      {{- if .Values.affinity }}
       affinity:
-        podAntiAffinity:
-          requiredDuringSchedulingIgnoredDuringExecution:
-            - labelSelector:
-                matchExpressions:
-                  - key: "app"
-                    operator: In
-                    values:
-                    - hfds-namenode
-              topologyKey: "kubernetes.io/hostname"
+{{ toYaml .Values.affinity | indent 8 }}
+      {{- end }}
+      {{- if .Values.nodeSelector }}
+      nodeSelector:
+{{ toYaml .Values.nodeSelector | indent 8 }}
+      {{- end }}
+      {{- if .Values.tolerations }}
+      tolerations:
+{{ toYaml .Values.tolerations | indent 8 }}
+      {{- end }}
       containers:
         # TODO: Support hadoop version as option.
         - name: hdfs-namenode
@@ -243,12 +245,6 @@ spec:
             - name: kerberos-keytab-copy
               mountPath: /kerberos-keytab-copy
       {{- end }}
-      {{- if .Values.namenodePinningEnabled }}
-      # Pin namenode pods to a set of nodes. You can label your node like below:
-      #   $ kubectl label nodes YOUR-NODE hdfs-namenode-selector=hdfs-namenode
-      nodeSelector:
-        hdfs-namenode-selector: hdfs-namenode
-      {{- end }}
       restartPolicy: Always
       volumes:
         - name: nn-start

charts/hdfs-namenode-k8s/values.yaml

@@ -39,9 +39,6 @@ zookeeperQuorum: zk-0.zk-svc.default.svc.cluster.local:2181,zk-1.zk-svc.default.
 # a standby namenode. Separated by the semicolon character.
 journalQuorum: hdfs-journalnode-0.hdfs-journalnode.default.svc.cluster.local:8485;hdfs-journalnode-1.hdfs-journalnode.default.svc.cluster.local:8485;hdfs-journalnode-2.hdfs-journalnode.default.svc.cluster.local:8485
 
-# Whether or not to enable pinning of namenode pods to labeled k8s cluster nodes.
-namenodePinningEnabled: false
-
 # Whether or not to use hostNetwork in namenode pods. Disabling this will break
 # data locality as namenode will see pod virtual IPs and fails to equate them with
 # cluster node physical IPs associated with data nodes.
@@ -87,3 +84,17 @@ kerberosConfigFileName: krb5.conf
 # where HOST-NAME should match the cluster node
 # host name that each per-host hdfs principal is associated with.
 kerberosKeytabsSecret: hdfs-kerberos-keytabs
+
+# Node labels and tolerations for pod assignment
+nodeSelector: {}
+tolerations: []
+affinity:
+  podAntiAffinity:
+    requiredDuringSchedulingIgnoredDuringExecution:
+      - labelSelector:
+          matchExpressions:
+            - key: "app"
+              operator: In
+              values:
+              - hfds-namenode
+        topologyKey: "kubernetes.io/hostname"

charts/hdfs-simple-namenode-k8s/templates/namenode-statefulset.yaml

@@ -42,6 +42,18 @@ spec:
       labels:
         app: hdfs-namenode
     spec:
+      {{- if .Values.affinity }}
+      affinity:
+{{ toYaml .Values.affinity | indent 8 }}
+      {{- end }}
+      {{- if .Values.nodeSelector }}
+      nodeSelector:
+{{ toYaml .Values.nodeSelector | indent 8 }}
+      {{- end }}
+      {{- if .Values.tolerations }}
+      tolerations:
+{{ toYaml .Values.tolerations | indent 8 }}
+      {{- end }}
       # Use hostNetwork so datanodes connect to namenode without going through an overlay network
       # like weave. Otherwise, namenode fails to see physical IP address of datanodes.
       hostNetwork: true
@@ -67,10 +79,6 @@ spec:
           volumeMounts:
             - name: hdfs-name
               mountPath: /hadoop/dfs/name
-      # Pin the pod to a node. You can label your node like below:
-      #   $ kubectl label nodes YOUR-NODE hdfs-namenode-selector=hdfs-namenode-0
-      nodeSelector:
-        hdfs-namenode-selector: hdfs-namenode-0
       restartPolicy: Always
       volumes:
         - name: hdfs-name

charts/hdfs-simple-namenode-k8s/values.yaml

@@ -29,3 +29,8 @@ nameNodeHostPath: /hdfs-name
 customHadoopConfig: {}
   # Set variables through a hash where env variable is the key, e.g.
   # HDFS_CONF_dfs_datanode_use_datanode_hostname: "false"
+
+# Node labels and tolerations for pod assignment
+nodeSelector: {}
+tolerations: []
+affinity: {}