@microscopic-river-74022 with persistentvolumeclaim actions going away, am I encouraged not to use the container action type either (since it is not compatible with persistent volumes anymore) or is there an alternative recommended way to persist data in the garden stack?

@microscopic-river-74022 I got this error when running

garden deploy --sync

, help me debug and fix it

Could not flush synchronization changes, retrying

yooo @microscopic-river-74022 can you help me figure out how to use the k8s provider's deploy action (edge/0.14) to change the namespace a kustomize overlay is pushed to

@microscopic-river-74022 : How can I replace a garden bonsai local mode configuration with reverse port forwarding with an equivalent sync configuration for garden 0.14?

@microscopic-river-74022 local.demo.garden isn't working for me but localhost:port is. I just get NS_ERROR_UNKNOWN_HOST when trying to connect to the service using local.demo.garden. I'm running on macOS. logs show

deploy.frontend           → Ingress: http://<project>.local.demo.garden

. I'm using garden version 0.14.4

@microscopic-river-74022 what should I put as imagepullsecrets and deploymentregistry when using ghcr.io with kubernetes provider?

@microscopic-river-74022 Why isn't

exec <pod> -i -- sh

working on garden dev console? It just says "Command exec completed successfully!".

garden exec <pod> -i -- sh

from the commandline works for the same pod.

@microscopic-river-74022 Why am I getting an error that says React is the wrong version?

@microscopic-river-74022 Why am I getting

Could not find key kubeconfig under var

error when running

garden validate

for the following project config?

apiVersion: garden.io/v2
kind: Project
name: pulumi
environments:
  - name: init
  - name: dev

providers:
  - name: pulumi
    version: null  # use whatever is in path
    useNewPulumiVarfileSchema: true
  - name: kubernetes
    buildMode: cluster-buildkit
    kubeconfig: ${var.kubeconfig}?  # use $KUBECONFIG if deployment doesn't set this explicitly

It should work according to docs, it's actually the example listed there: https://docs.garden.io/features/variables-and-templating#optional-values

@microscopic-river-74022 How can I make

kubeconfig

optional for the kubernetes provider with garden 0.14? When I set it to

kubeconfig: ${var.kubeconfig || null}

I get the following error:

✖ providers                 → Failed processing resolve provider kubernetes (took 0.15 sec). This is what happened:

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Error validating provider configuration:

kubeconfig must be a string
────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
✔ pulumi                    → Provider is ready
Failed resolving one or more providers:
- kubernetes

@microscopic-river-74022 Can I somehow use outputs of a one garden project as inputs for another garden project?

@microscopic-river-74022 Why is this provider config

providers:
  - name: pulumi
    version: null  # use whatever is in path
    useNewPulumiVarfileSchema: true
  - name: exec
    preInit:
      runScript: mise run get kubecontext
  - name: kubernetes
    dependencies:
      - exec
    preInit:
      runScript: mise run write-kubeconfig
    buildMode: cluster-buildkit
    kubeconfig: ./infra/kubeconfig.yaml
    context: ${providers.exec.outputs.initScript.log}

giving me the following error?

❯ garden validate --force-refresh
Validate ✔️ 

ℹ garden                    → Initializing...
ℹ garden                    → Running in environment dev.default in project pulumi
ℹ providers                 → Resolving providers (will force refresh statuses)...
ℹ exec                      → Configuring provider...
ℹ pulumi                    → Configuring provider...
✔ exec                      → Provider configured
✔ pulumi                    → Provider configured
ℹ exec                      → Running pre-init script
✔ pulumi                    → Provider is ready
ℹ exec                      → kubernetes-admin@huudot-dev-cluster
ℹ exec                      → Finished in 28.0ms
ℹ exec                      → Pre-init script completed successfully
✔ exec                      → Provider is ready
✖ providers                 → Failed processing resolve provider kubernetes (took 0.25 sec). This is what happened:

 
project.garden.yml:22
...
21  |     kubeconfig: ./infra/kubeconfig.yaml
22  |     context: ${providers.exec.outputs.initScript.log}
-------------------^
Could not find key initScript under providers.exec.outputs. Available keys: (none).
 
Failed resolving one or more providers:
- kubernetes

@microscopic-river-74022 Why am I getting the following error when trying to deploy a helm chart using

garden deploy

?

ℹ deploy.tailscale-operator → Deploying tailscale-operator (type helm) at version v-9f42f2e248...
✖ deploy.tailscale-operator → Failed (took 6.3 sec)
✖ deploy.tailscale-operator → Failed processing Deploy type=helm name=tailscale-operator (took 8.23 sec). This is what happened:

─────────────────────────────────────────  
Command "/Users/spike/.garden/tools/helm/0dc3c32729046d4e/darwin-arm64/helm --kube-context kubernetes-admin@huudot-test-cluster --kubeconfig /Users/spike/gits/work/huudot.git/main/iac/infra/kubeconfig.yaml --namespace tailscale install tailscale-operator tailscale-operator --repo https://pkgs.tailscale.com/helmcharts --dry-run --namespace tailscale --output json --timeout 300s --values /private/var/folders/1y/httmdjmd02qf57f9zd3py6kw0000gn/T/113b7979b6cec659e9bc40edeee803a5" failed with code 1:

Error: INSTALLATION FAILED: file '/Users/spike/Library/Caches/helm/repository/tailscale-operator-1.86.2-1753978303-9f1ee41e298aac17db343df700f412b84d349962fbb0f680b102fa5e2c1453fd.tgz' does not appear to be a gzipped archive; got 'application/octet-stream'

@microscopic-river-74022 I have services which use garden + terraform. For some reason they randomly have their secrets wiped out which are stored in secrets manager. I intialy create a secret in secret manager using terraform with an intial value later I manully add more values to it and thus the service can use it but randomly not everytime they get wiped out and get reset to the intial value I have the below block to prevent this but its still happening any suggestions ?

resource "aws_secretsmanager_secret_version" "app_secret_version" {
  secret_id     = aws_secretsmanager_secret.app_secret.id
  secret_string = var.secret_string

  lifecycle {
    ignore_changes = all
  }
}

@User I am using garden deploy action with the kustomize manifest option. When I change my manifest how do i direct garden to redeploy ONLY the manifest changes and not perform a new rebuild of the container image since no application code was changed.

@microscopic-river-74022 how would run a third party tool within garden (trivy for example) to scan images before they are published using

garden publish

? Preferably scan before the push occurs.

@microscopic-river-74022 is it possible run 2 garden commands sequentially using a 1 custom command (https://docs.garden.io/features/custom-commands). I would like to run cleanup deploy and cleanup namespace as an optional custom command. I tried something like this and it does not work:

kind: Command
name: cleanup-all
description:
  short: Cleans up all environments (preview in the current project configuration)
gardenCommand:
  - cleanup
  - deploy
  - app,app2
  - "&&"
  - cleanup
  - namespace

The other option might be to create 2 custom commands and then call that command (similar to a Makefile target), but this would require 3 custom commands. I am trying to avoid that.

@microscopic-river-74022 , I am trying to display project outputs after running a Build action. The reference in project.garden.yaml is

outputs:
  - name: app_image
    value: "${actions.build.app.outputs.deploymentImageName}"

I have a Build Action with the name

app

defined. I run

garden build --env ci

which completes successfully running the container build. I then run `` and receive the following error:

57  |   - name: app_image
58  |     value: "${actions.build.app.outputs.deploymentImageName}"
-----------------^
Could not find key actions. Available keys: local, command, datetime, project, git, secrets, variables, var, environment, providers, modules, runtime.

@microscopic-river-74022 suggest how to deploy a configmap with templated values based on project variables

@microscopic-river-74022 Is there any way to prevent garden from printing the task output to logs? In case the output contains some sensitive info like secrets.

@microscopic-river-74022 Is there a way to wait for a CRD to reach a certain status before deploying something that depends on it? The CRD (cnpg cluster to be exact) gets deployed by garden but putting that deployment as a dependency only waits until the deployment is done, not until the CRD is ready.

@microscopic-river-74022 Is there a way to make cli tell what it is doing or waiting for during deployment? It's not always clear what is going on from the logs

@microscopic-river-74022 Is there a way to provide helm deploy action with arbitrary cli flags? In particular, I'm trying to get

traefik

deployed without bundled crds. This is how you do it when installing the helm chart manually:

helm install traefik traefik/traefik --skip-crds

@microscopic-river-74022 is there a way to output all the current outputs.deploymentImageId for every build action in the project.

@microscopic-river-74022 Is there a way to deploy kubernetes manifest files from a url? Trying to deploy Gateway API resources: https://github.com/kubernetes-sigs/gateway-api/releases/download/v1.4.0/standard-install.yaml

@microscopic-river-74022 Can I split the

providers.exec.outputs.initScript.log

and use the last line as context for the kubernetes provider? Something like this:

context: ${split("${providers.exec.outputs.initScript.log}", "\n")[-1]}

Please stand by @ancient-island-74119 as I search the knowledge base. I normally take about 15 seconds or so. :wave:

@microscopic-river-74022 Can I somehow easily see what contributes to the version calculation of an action?

@microscopic-river-74022 Can I use negation in the exclude field? So that the specified file would not be excluded?

@microscopic-river-74022 Why is deploy and run actions running on consecutive `garden deploy`s even though according to logs the action version is identical on both times? Shouldn't

garden

skip the actions that haven't changed since the last run?