#helmfile - January 2021

Quick little utility program I came up with to convert Kubernetes YAML to a format that can be used by the incubator/raw helm chart: https://github.com/RothAndrew/convert-to-helm-raw-values

Hi everyone,
Does anyone have a problem with helmfile concurrency? I'm not able to install multiple releases in parallel, I see only one helm upgrade/install process at the time. Latest helmfile, tried on Windows (exe) and also tried to run as docker image.

Happy New Year @mumoshu, hope you are well! Just wanted to get your attention on a vals PR when you have time: https://github.com/variantdev/vals/pull/40 Hoping to get this in Helmfile after merge too. Please let me know if there is a different maintainer I should ping instead.

Hello there! Are we able to import releases through http with helmfile? I'd like to use releases from cloudposse/helmfiles without using submodules/cloning the whole repo

Guys? How are You using secrets in helmfile ? Currently I am sourcing bash script, which will via azure cli get proper secrets and then I am using requiredEnv in helmfile...
Now I have these problems:
1. argocd can install only binary plugins, so helm-secret will not help me, unless i customize my own argocd image, which is no go ( sops with azure keyvault )
2. i am not sure what exactly is variantdev/vals and if it can replace helm-secret plugin
3. seems that helm file has no integration to azure keyvault
Any idea? Thanks

Hey guys, thought I’d finally join the Slack channel since I’ve being working with Helmfile for a fair while now. I’ve just come back to work after a short holiday and I’m facing a strange issue after the latest update if anyone has any ideas what could be causing it?

in ./helmfile-3dce096a66df5e8515ae85f40bd83a2bc10620b819f7b5c0019505065acd204d.yaml: in .helmfiles[1]: in helmfile/releases/01-misc-helmfile.yaml: 2 errors:
10:44:31  err 0: failed processing release nfs-server-qa: open .terraform/helmfile/temp-7847bc44b4/qa-nfs-server-qa-values-855cb5cbd4: no such file or directory
10:44:31  err 1: failed processing release redis-tasks-qa: open .terraform/helmfile/temp-7847bc44b4/qa-redis-tasks-qa-values-5c78bf6c78: no such file or directory

This happens on any charts, the only one it does not happen on is rbac chart but that is only because it never loads a values file. Anyone faced this before?

^ Continuing on from above it would seem if I revert back to the previous version of the helmfile provider 0.11.1 I don’t experience this issue so I’m guess it has something to do with the --skip-diff-on-install @mumoshu

This is using 14.4 terraform and latest versions of both helmfile provider and helmfile

Hey everyone ! Does helmfile status showing STATUS: deployed guarantee that the pods are in running state or it shows deployed as soon as they are scheduled ?

Any idea if is possible to include raw kubeconfig as variable, instead of passing kubeconfig file in https://github.com/mumoshu/terraform-provider-helmfile ? For example with this provider https://github.com/vmware-tanzu/terraform-provider-carvel/blob/develop/pkg/provider/kubeconfig.go#L33 is possible to reference kubeconfig as variable without need of existing kubeconfig file ...

Hello! I am very thankful for the work on helmfile, which has been a lot of fun to use!

A bit of a strange question here - is it possible to "pass" variables / outputs between releases? I.e. I have a helm release that depends on dynamic values from a previous helm/helmfile release (i.e. in a different directory) - I could get the values by shelling out to kubectl or something like that, but that feels really dirty. Wondering if anyone has ideas for a better way 🙂

hello all,
can somebody give me an example about strategicMergePatches?

How do I escape values that have $ in the values?
I am getting:

in ./helmfile.yaml: failed processing release my-prometheus: failed to render values files "dgraph-app-alert-rules.yaml.gotmpl": failed to render [dgraph-app-alert-rules.yaml.gotmpl], because of template: stringTemplate:11: undefined variable "$labels"

Hi, all!

Do you know if it’s possible to have helmfiles in helmfiles? 2 levels

I am unsure what to do with this, as I don't use environments for this mini-project:

 0: --
 1: repositories:
 2:   - name: prometheus-community
 3:     url: <https://prometheus-community.github.io/helm-charts>
 4:   - name: stable
 5:     url: <https://charts.helm.sh/stable>
 6: 
 7: releases:
 8:   - name: my-prometheus
 9:     namespace: monitoring
10:     chart: prometheus-community/kube-prometheus-stack
11:     values:
12:       - ./dgraph-prometheus-operator.yaml
13:       - grafana:
14:           adminPassword: admin123
15:      disableValidation: true
16: 

could not deduce `environment:` block, configuring only .Environment.Name. error: failed to read helmfile.yaml.part.0: reading document at index 1: yaml: line 2: mapping values are not allowed in this context

Is there something obviously wrong in yaml that I missed?

when I use mergepatch and try to deploy patched version of chart to my cluster it quits with error it does not find secret,configmap,statefulset etc... every time it created the mentioned resources but not in the time when it checked. it creates the release secret so next time I cannot install it again it sees everything was fine during install. If I remove the release secret file I can run helmfile apply again and it goes 1 step further and exit with different resource is missing error however those resources are in the patched yaml... Any idea ?

how can i limit helmfile to a specific release?

For remote helmfiles (helmfiles referenced by git path), do people keep those helmfiles in separate repositories for easy versioning? Similar to how people generally do it with terraform modules?

Hi, I'm new with helmfile and having a problem with the needs keyword. I have a very simple file which only applies cert-manager and my cluster issuers. When i run helmfile applyit errors out by saying:

no matches for kind "ClusterIssuer" in version "cert-manager.io/v1alpha2"

I therefore use the needs keyword to make sure that helmfile installs cert-manager before my issuers but that is not working. If I comment out the issuers and first install cert-manager and then uncomment issuers and install them everything works. I guess I'm at fault here and that it probably is an obvious error but I can't see what I'm doing wrong?

repositories:
  - name: prometheus-community
    url: <https://prometheus-community.github.io/helm-charts>
  - name: stable
    url: <https://charts.helm.sh/stable>

helmDefaults:
  createNamespace: true

releases:
  - name: cert-manager
    namespace: cert-manager
    chart: jetstack/cert-manager
    wait: true
    set:
      - name: installCRDs
        value: true
    version: v1.1.0
  
  - name: issuers
    namespace: cert-manager
    chart: issuers/
    needs:
      - cert-manager/cert-manager
    secrets:
      - issuers/secrets.yaml

Hi there! I'm trying to set missingFileHandler: Warn for values: section of my helmfile:

bases:
- ../_lib/defaults.yaml.gotmpl
---
{{ readFile "../_lib/templates.yaml.gotmpl" }}

values:
- hosts: []
#  - name: "example"
#    ingressClass: "nginx"
#    port: "80"
#    targetPort: "80"
#    upstreamAddress: "192.168.0.1"
#    upstreamVhost: "<http://www.example.com|www.example.com>"
#    hostname: "<http://www.example.com|www.example.com>"
- ../../config/dbcc-external-services/{{ .Environment.Name }}/values.yaml.gotmpl        # Import env-specific list of hosts

releases:
- name: dbcc-external-services
  <<: *dbcc
  chart: incubator/raw
  namespace: miscellaneous-extra
  version: 0.2.3

I cannot use Environments: here, because it's complimentary helmfile included in a few entry-level helmfiles with different sets of environments...
Any advice is much appreciated!

Hello, are there any examples of how to consume https://github.com/cloudposse/helmfiles/tree/master/releases ? is the design to consume these in ways such as

helmfiles:
- 
  path: git::<https://github.com/cloudposse/helmfiles.git@releases/echo-server/helmfile.yaml?ref=0.125.0>

and if so, how do I have more environments than specified in the echo-server/environments.yaml ?

hi, noob here, do you know a howto i can follow with this task i wanted to do? basically i want to create a postgres pod using helm possibly and point it to a persistent drive.. i am not sure if there is a way to pre create the databases as well, everytime a pod with postgres is created? please help point out a resource that i can read

hello all , I have a small question on exec template function. I want to execute split command and pass some vars to this command to be able to grab the values out of this , but I am unsure if this exec is for executing the scripts or I can just use any commands out of this , how does this internally execute this (like if in the shell this already runs in bash shell) ?

Does anyone know if the needs: keyword works with separate helmfiles (i.e. in a helmfile.d/ directory)? I'd like to break my helmfiles up, but I don't really like the idea of using the 00- , 01- , etc prefixes for each file in order to get them to apply in a specific order.

Hi all
helmfile supports reading values from Hashicorp Vault via ref tool,
but does it support retrieving values from Consul ?

I'm seeing an error where i'm trying to install kube-prometheus-stack using helmfile but it appears to not be installing the stuff it depends on (CRDs specifically). does helmfile expect me to put these in manually or is it likely that im doing something wrong and helmfile would normally install all the chart's dependencies? using helm3 and helmfile v0.137.0

I’ve been using ArgoCD and helmfile together and as suggested in the README, I do a helmfile template in my CI to achieve gitops for my apps.

So question, do people also manage their base infrastructure charts (cert-manager, nginx/traefik chart, etc) in Argo? Or still let helmfile manage them? A bit weird to have two ways to deploy to the cluster but I guess that’s ok.

hello all, Is there a way to get info for helmfile from cluster before we do an apply ? Currently we provide values with env from OS, but it would be nice if we could grab it from k8s.