Added rollout strategy override and automation

Rollout strategy for deployments is now volume and autoscaler aware.
It will now default to `Recreate` when one or more volumes and a replica of 1 is
found with no autoscaler. This ensures rollouts can still occur when
there is only one replica which would otherwise stall.

Improvement would be to also add volume permission detection, so this
only happens when RWO is the only setting. Since RWX would still work
even with non Recreate strategy.

charts/corvid-test/Chart.yaml

@@ -3,8 +3,8 @@ appVersion: 0.1.0
 description: A Helm chart for Kubernetes
 name: corvid-test
 type: application
-version: 0.7.0
+version: 0.8.0
 dependencies:
 - name: corvid
-  version: 0.11.0
+  version: 0.12.0
   repository: "file://../corvid"

charts/corvid-test/README.md

@@ -2,7 +2,7 @@
 
 A Helm chart for Kubernetes
 
-![Version: 0.7.0](https://img.shields.io/badge/Version-0.7.0-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 0.1.0](https://img.shields.io/badge/AppVersion-0.1.0-informational?style=flat-square)
+![Version: 0.8.0](https://img.shields.io/badge/Version-0.8.0-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 0.1.0](https://img.shields.io/badge/AppVersion-0.1.0-informational?style=flat-square)
 
 ## Installing the Chart
 
@@ -18,7 +18,7 @@ $ helm install corvid-test raven/corvid-test
 
 | Repository | Name | Version |
 |------------|------|---------|
-| file://../corvid | corvid | 0.11.0 |
+| file://../corvid | corvid | 0.12.0 |
 
 ## Values
 
@@ -34,6 +34,7 @@ $ helm install corvid-test raven/corvid-test
 | cron.enabled | bool | `false` | enable or disable cronjob |
 | cron.schedule | string | `"@midnight"` | schedule for cronjob using Cron syntax https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/#schedule-syntax |
 | cron.suspend | bool | `false` | cronjob will not trigger on schedule but can be manually triggered |
+| deployment.strategy | string | `""` | rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume |
 | dnsConfig | object | `{}` |  |
 | dnsPolicy | string | `""` |  |
 | envFrom | string | `nil` |  |

charts/corvid-test/values.yaml

@@ -26,6 +26,10 @@ command:
 args:
 runtimeClassName:
 
+deployment:
+  # -- rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume
+  strategy: ""
+
 cron:
   # -- enable or disable cronjob
   enabled: false

charts/corvid/CHANGELOG.md

@@ -1,5 +1,27 @@
 # Changelog
 
+## 0.12.0
+
+Behaviour has changed for single replica deployments with volumes.
+If you have a deployment that:
+
+- has a single replica (any of .Values.replicaCount=1, .Values.autoscaling.enabled=false)
+- has a volume (any of .Values.volumes>0, .Values.persistence.existingClaim!="", .Values.persistence.enabled=true)
+
+We now default the deployment to `Recreate` rollout strategy instead of `RollingUpdate` to enable rollouts to complete without manual intervention.
+
+You can manually define the rollout strategy if you would rather not rely on this automated behaviour, this is left blank by default but feel free to adjust it as you see fit downstream:
+
+```yaml
+
+deployment:
+  # -- rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume
+  strategy: ""
+
+```
+
+We do check for this key, so you will need to add this to your `values.yaml` else you will get thrown a key error.
+
 ## 0.11.0
 
 This is backwards incompatible.

charts/corvid/Chart.yaml

@@ -1,6 +1,6 @@
 apiVersion: v2
 appVersion: "0.1.0"
-version: 0.11.0
+version: 0.12.0
 description: Common helm component and utility library
 name: corvid
 type: library

charts/corvid/README.md

@@ -2,7 +2,7 @@
 
 Common helm component and utility library
 
-![Version: 0.11.0](https://img.shields.io/badge/Version-0.11.0-informational?style=flat-square) ![Type: library](https://img.shields.io/badge/Type-library-informational?style=flat-square) ![AppVersion: 0.1.0](https://img.shields.io/badge/AppVersion-0.1.0-informational?style=flat-square)
+![Version: 0.12.0](https://img.shields.io/badge/Version-0.12.0-informational?style=flat-square) ![Type: library](https://img.shields.io/badge/Type-library-informational?style=flat-square) ![AppVersion: 0.1.0](https://img.shields.io/badge/AppVersion-0.1.0-informational?style=flat-square)
 
 This library chart primarily deals with abstracting common boilerplate into customisable components for re-use.
 
@@ -30,6 +30,7 @@ $ helm install corvid raven/corvid
 | cron.enabled | bool | `false` | enable or disable cronjob |
 | cron.schedule | string | `"@midnight"` | schedule for cronjob using Cron syntax https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/#schedule-syntax |
 | cron.suspend | bool | `false` | cronjob will not trigger on schedule but can be manually triggered |
+| deployment.strategy | string | `""` | rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume |
 | envFrom[0].configMapRef.name | string | `"someConfigMap"` |  |
 | envFrom[0].configMapRef.optional | bool | `false` |  |
 | envFrom[1].secretRef.name | string | `"someSecret"` |  |
@@ -95,6 +96,28 @@ $ helm install corvid raven/corvid
 
 # Changelog
 
+## 0.12.0
+
+Behaviour has changed for single replica deployments with volumes.
+If you have a deployment that:
+
+- has a single replica (any of .Values.replicaCount=1, .Values.autoscaling.enabled=false)
+- has a volume (any of .Values.volumes>0, .Values.persistence.existingClaim!="", .Values.persistence.enabled=true)
+
+We now default the deployment to `Recreate` rollout strategy instead of `RollingUpdate` to enable rollouts to complete without manual intervention.
+
+You can manually define the rollout strategy if you would rather not rely on this automated behaviour, this is left blank by default but feel free to adjust it as you see fit downstream:
+
+```yaml
+
+deployment:
+  # -- rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume
+  strategy: ""
+
+```
+
+We do check for this key, so you will need to add this to your `values.yaml` else you will get thrown a key error.
+
 ## 0.11.0
 
 This is backwards incompatible.

charts/corvid/templates/_deploy.yaml

@@ -1,4 +1,8 @@
 {{- define "corvid.deploy.tpl" -}}
+{{- $hasVolume := (or (ne (len .Values.volumes) 0) (or (ne .Values.persistence.existingClaim "") (.Values.persistence.enabled))) }}
+{{- $hasOneReplica := (and (eq (int .Values.replicaCount) 1) (not .Values.autoscaling.enabled)) }}
+# has volume: {{ $hasVolume }}
+# has 1 replica: {{ $hasOneReplica }}
 apiVersion: apps/v1
 kind: Deployment
 metadata:
@@ -9,6 +13,13 @@ spec:
   {{- if not .Values.autoscaling.enabled }}
   replicas: {{ .Values.replicaCount }}
   {{- end }}
+  {{- if .Values.deployment.strategy }}
+  strategy:
+    type: {{ .Values.deployment.strategy }}
+  {{- else if and $hasVolume $hasOneReplica }}
+  strategy:
+    type: "Recreate" # detected a single replica with a volume which would block on rollout if Rollout was used
+  {{- end }}
   selector:
     matchLabels:
       {{- include "corvid.selectorLabels" . | nindent 6 }}

charts/corvid/values.yaml

@@ -25,6 +25,10 @@ command:
 args:
 runtimeClassName:
 
+deployment:
+  # -- rollout strategy `Recreate` or `RollingUpdate` this chart defaults to Recreate only if we detect a single replica with a volume
+  strategy: ""
+
 cron:
   # -- enable or disable cronjob
   enabled: false