He guys, when I google I get some results on dagger.Chart or even helm.Chart but when I check up on https://dagger-io.readthedocs.io/en/sdk-python-v0.6.4/ it does not seem to be supported at all to install helm charts (in my local k3d cluster) using dagger. Can someone point me to some working examples?
#Helm support in dagger python-sdk?
rocky charm
shut venture
hey Emiel! you're probably getting old indexed results from the CUE SDK which had a sort of universe concepts with different packages.
since dagger still doesn't support container-to-host networking (#general message), connecting to your local kubernetes cluster is currently possible but might require some hacks to get it working. Are you using a linux machine?
rocky charm
Yeah, I'm using an Ubuntu VM. Using google I came across some code using dagger.Chart and dagger.Helm. These don't seem to be available (anymore) in the latest SDK. Also seen some messages indicating that it was indeed deprecated. Seems I will have to do it through subprocess.run statements. Pity, was hoping dagger would have some nice abstraction layer around helm installs to keep my pipeline code relatively simple.
PS: I just want dagger to be able to point a helm install to some remote host. The network connectivity part is not what I am worried about.
I am only testing it locally on my k3d cluster because I do not have my access to my 'real' testcluster yet.
shut venture
Seems I will have to do it through subprocess.run statements. Pity, was hoping dagger would have some nice abstraction layer around helm installs to keep my pipeline code relatively simple.
Dagger has always been about running things in containers. The CUE dagger.Chart and dagger.Helm very likely just ran the alpine/helm docker container for that. You don't need to use subprocess.run for your pipeline, you can directly use the alpine/helm container.
Something like.
client.from_("alpine/helm").with_exec("helm", "install", "mychart")
rocky charm
Thanks, when I get around to it again I will try this approach then.