* Add configurable session affinity
* Bump version of Consul chart
* Update README with new value
* Added sessionAffinity to values-production.yaml
* Align use of spaces in consul-service.yaml
* fix: adding "-binding" suffix to the PVC claimName
"-binding" suffix is missing from the deployment volume declaration, while it is present in the PVC description. Adding the suffix fixes the persistency feature of this chart, that is otherwise broken.
* feat: update chart version
* [bitnami/consul] fix: use the service DNS instead pod DNS for join the cluster
Signed-off-by: darteaga <darteaga@vmware.com>
* fix: remove unnecessary variables
Signed-off-by: darteaga <darteaga@vmware.com>
* [bitnami/metallb] Fixed the PrometheusRule not creating
* [bitnami/metallb] prometheus.enabled parameter added for creating PrometheusRules
* [bitnami/metallb] prometheus.enabled default value setted to false
* [bitnami/metallb] version bumped
* Sync back most significant upstream projectcontour changes
The most important change is the refactoring of the shutdown sidecar:
7cd9f4a685
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Delete removed CRD ingressroutes.contour.heptio.com
> As a reminder, support for IngressRoute was officially dropped in v1.6.
> If you haven’t already migrated to HTTPProxy, see the IngressRoute to
> HTTPProxy migration guide for instructions on how to do so. Once you have
> migrated, delete the IngressRoute and related CRDs.
Ref: https://projectcontour.io/resources/upgrading/
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Sync HTTPProxy CRD
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Sync TLSCertificateDelegation CRD
* Drop old TLSCertificateDelegation CRD on API group contour.heptio.com
API group contour.heptio.com is no longer supported since 1.6.x
* Bump minor chart version as the change is no longer a patch
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Attempt to fix contour tests by upgrading to latest github actions
* Override envoy service type to ClusterIP to make tests happy
I found a hint in PR #2721 that it should be possible to override values
used in GH actions.
* Revert "Attempt to fix contour tests by upgrading to latest github actions"
This was requested during the review process by dani8art.
This reverts commit 253a8ecd60.
* Add bitnami/common as a dependency
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Also use bitnami/common subchart for image
* Implement extraVolumes and extraVolumeMounts on contour and envoy
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Implement extraEnvVars on contour and envoy
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Implement extraEnvVarsConfigMap and extraEnvVarsSecret
* Implement initContainers for contour and envoy
* Implement service.extraPorts on contour and envoy
* Implement rolling tags helpers
* Sync rbac with upstream
As George Goh (@georgegoh) mentioned in issue projectcontour/contour#2050
the leaderelection role is not needed anymore:
> 1. I remove the `contour-leaderelection` role as it's no longer needed
> in `templates/rbac.yaml`.
> 2. General matching of the rbac.yaml to the rbacs defined in this project.
Signed-off-by: Marco Kilchhofer <marco@kilchhofer.info>
* Update README.md
* Fix default values for image repositories
* Add quotes (`)
* Fix whitespaces in table
* Also use bitnami/common subchart for pullSecrets
The subchart bitnami/common now supports this (version 0.6.2 and newer).
See PR #3566
* Use same tpl functions for job
Inside the Deployment we can use templating with 'tpl'. Since we use the
same structure (affinity, nodeSelector and tolerations) for the
Deployment and the Job, we should also use the same template functions.
* Fix: affinity defined twice
* Use more specific keyword antiAffinity -> antiAffinityPolicy
* Bump chart major: 2.0.0
* Variant2: Implement certgen by using hooks (#2)
* Implement certgen via hooks
* Delete resources with helm hooks only if needed
* Revert "Delete resources with helm hooks only if needed"
This reverts commit 76449252d06b7bda0f16c490316478c3fb1004f1.
As documented inside the helm docs, we should remove unneeded resources:
~~~
The resources that a hook creates are currently not tracked or managed as part
of the release. Once Helm verifies that the hook has reached its ready state,
it will leave the hook resource alone. Garbage collection of hook resources
when the corresponding release is deleted may be added to Helm 3 in the future,
so any hook resources that must never be deleted should be annotated with
`helm.sh/resource-policy: keep`.
Practically speaking, this means that if you create resources in a hook, you
cannot rely upon helm uninstall to remove the resources. To destroy such
resources, you need to either add a custom `helm.sh/hook-delete-policy`
annotation to the hook template file, or set the time to live (TTL) field of
a Job resource.
~~~
* Replace colons with dashes (standardize even more)
This was requested during review by @dani8art
* Clearify README regarding CRDs and helm v3
* Add small upgrading notes to README
* Fix helm2 incompatibility due to PR #2961
Helm2 uses Sprig v2 and therefore the funtion "get" is not available
there. Since we need to guarantee helm v2 support, we need to workaround
this.
* Use consistent component labels on certgen resources
* Use bitnami/common for apiVersion of kind Deployment
* Rename parameter `contour.createCustomResource` to `contour.installCRDs`
* Do not allocate an IP address on metrics services
Services for ServiceMonitor do not need to allocate an IP address. We therefore do not
waste IPs from the services CIDR pool for this.
* Split into subfolders
