#helmfile - February 2021

Can somebody provide an example for helmfile with vals and aws secret store ? Thanks

Chart incubator/raw was marked as deprecated, part of helm charts repository deprecation (https://github.com/helm/charts/tree/master/incubator/raw). Is there a new home/repo for it or anyone step up for maintaining it? Or is there a better way define resources via values? I’m using this chart in helmfile to define CRs, so they nicely coupled with other releases, for example cert-manager deployment and ClusterIssuer defined in single helmfile.

Helmfile w/o environments ? I have a task to simplify our helmfile, and let only run with a simple helmfile apply. Currently we provide cluster related data (fqdn, cert arn etcc) using env. I have few ideas:
• using vals with secret manager and grab all data from there wit ref. Can get all data from here
• using hashicorp vault inside cluster and dynamically use secrets. get cluster related info with exec...
• using randAlpha to generate missing secrets and use exists if there is any. get cluster related info with exec.
Any advice what else I should check and try ?

Is it possible, to mark a release so that it only runs when a --selector name=whatever is applied? wanting to do https://github.com/roboll/helmfile/tree/master/examples#managing-oneshot-jobs-with-helmfile this, but our CI runs helmfile apply on changes, and I really only want those jobs to run when explicitly told to do so.

would something like that work? Where it is false, unless ran with ENABLE_MIGRATIONS=true ?

i assume no 😕

I assume this means, we cant, not run these jobs when someone does a sync

Hi all
Is it a possible to share files that later could be used as bases in helmfile.yaml
For instance put them in Git or S3 or HTTP
And then fetch them
I've tried to use go-getter style path, but obtained an error

Is it possible to set helmDefaults in environments.yaml?

I'm looking to apply a global --wait=true , but i execute individual helmfiles in my environment and don't want to add a line to every file i'd rather add it in the exec of helmfile

is it maybe possible using the --args flag in helmfile ?

I see https://github.com/roboll/helmfile/issues/64 is still open

thats a bummer. just tried using --args "--wait", and it throws errors because it applies to every helm command like helm list

hello all,
I would like to add the Environment.Name as label to all alerts so alter I could setup a rule based on that. If I do like this helmfile does not catches it and in configmap I see without any changes.

I am seeking Helmfile advice:

We are deploying some Helm charts that are infra related.
At the moment we have only one cloud provider and use environments called stage and prod
A new cloud provider will be added next and we want to install the same set of helm charts there as well. The new cloud provider has some values that are only relevant to him.

How ca we best support that with helm?
I see this two options:
1. Add more envs eg. stage-aws, stage-azr, prod-aws, prod-azr and so on.
2. Continue using stage/prod but move to cloud provider specific stuff to selectors. And then do if selector aws do aws specific and so on
Do you see any addition option how to accomplish this?

Hey All, have a weird error. I'm providing a values file via a helmfile release to a helm chart. I'm providing it via a variable containing the full file path (the file is generated at run-time). The path I'm providing should be absolutely but is erroring because it doesn't exist in the relative .

values file matching "/tmp/manage-secrets-IhNjd4/combined_secrets.yaml " does not exist in "."

Has anyone ever seen anything like this?

perhaps this is not a helmfile question, but i think it might be? i know there's https://github.com/hashicorp/terraform-provider-aws/issues/16542#issuecomment-782606500 which is making nice progress but looking at the pr there's something that's missing for me.

i want to run coredns using a nodeSelector and coredns comes as a preinstalled Deployment with eks and/or an addon. i really don't want to overwrite the coredns deployment with my own deployment.. i'd rather just patch the existing deployment with a nodeSelector. i don't know any simple way to do this with helmfile?

I was wondering how a “selector” and “needs” works together. I have one “mongodb” chart and 2 “apps” that depend on it, Apps are labeled aith the selector “app=app1/2”. The apps have “need” setting to the mongo chart. When I use a selector to install app1, does it also install mongodb?

is there a standard way of "injecting" a template into helm charts before it's parsed with helmfile? i started playing with hooks looking for the right way but the working dir doesn't seem to be current dir. I can think of a number of cases where I might want to add a CRD or something to a chart but not want to have to keep up our own custom downstream version of the chart just to keep that one file in there. I see that there is a way to get kustomize to run with the chart but it would be nice to use all the _helpers in the chart i'm working on.

is anyone successfully using the remote environment values from 1296? i'm on helmfile v0.138.4 and it just treats the URLs as a local path

Hey has anyone made use of jsonPatches in helmfile? I am trying a very simple use-case for them when running a release against a directory of manifest files and I cannot seem to get it to work. A strategic merge patch works just fine but I wanted the ability to target and apply specifically

This is incredibly frustrating. I cannot get jsonPatches to pick up the target or properly modify the manifest being ingested. strategicMergePatch works, but it fails under my go-templating which iterates over many regions, the metadata/labels only get the region of the first iterated release