pulumi destroy

deletes everything in your stack, if that’s what you want

I’m not sure what’s going on with the provider (will open an issue)

it often cant

why can’t it?

what happens?

not sure, need to scroll. right now it decided to work so my console is blocked 🙂

it should always work 😛

well, in ideal world...

specifically, i meant that if it doesn’t, you should file bugs on us

we designed it for the exact purpose you’re describing of “clean slate”

ok, give me 30 seconds 😉

no worries!

we just want

pulumi destroy

to work since it’s really important 😛

for example, I've deleted the kubernetes cluster that was used by stack (I have 2 "phase" stack. first phase installs kubernetes, then I can run second phase to create application copies inside kubernetes), but the application stack is not aware of that and when it tries to do the deletion it just throws:

> pulumi destroy
Previewing destroy (feature-server-deployment):

     Type                               Name                                           Plan
 -   pulumi:pulumi:Stack                azure-py-kubernetes-feature-server-deployment  delete
 -   ├─ kubernetes:core:Secret          storageacct                                    delete
 -   ├─ kubernetes:core:Namespace       featureserverdeployment                        delete
 -   ├─ pulumi:providers:kubernetes     application_provider                           delete
 -   ├─ azure:signalr:Service           signalr                                        delete
 -   ├─ azure:storage:Account           storage                                        delete
 -   ├─ kubernetes:core:ServiceAccount  orleans-rbac                                   delete
 -   └─ azure:core:ResourceGroup        rg-branch                                      delete

Resources:
    - 8 to delete

Do you want to perform this destroy? yes
Destroying (feature-server-deployment):

     Type                       Name                                           Status                  Info
     pulumi:pulumi:Stack        azure-py-kubernetes-feature-server-deployment                          3 messages
 -   └─ kubernetes:core:Secret  storageacct                                    **deleting failed**     1 error

Diagnostics:
  kubernetes:core:Secret (storageacct):
    error: Plan apply failed: the cache has not been filled yet

  pulumi:pulumi:Stack (azure-py-kubernetes-feature-server-deployment):
    E0208 23:32:34.891272    8220 memcache.go:126] couldn't get current server API group list; will keep using cached value. (Get <https://dns-8920cff4.hcp.westeurope.azmk8s.io:443/api?timeout=32s>: dial tcp: lookup <http://dns-8920cff4.hcp.westeurope.azmk8s.io|dns-8920cff4.hcp.westeurope.azmk8s.io>: no such host)
    warning: Cluster failed to report its version number; falling back to 1.9%!(EXTRA bool=false)
    E0208 23:32:35.124311    8220 memcache.go:126] couldn't get current server API group list; will keep using cached value. (Get <https://dns-8920cff4.hcp.westeurope.azmk8s.io:443/api?timeout=32s>: dial tcp: lookup <http://dns-8920cff4.hcp.westeurope.azmk8s.io|dns-8920cff4.hcp.westeurope.azmk8s.io>: no such host)

Permalink: <https://app.pulumi.com/4c74356b41/azure-py-kubernetes/feature-server-deployment/updates/37>
error: update failed

I thought I had another case, but appears I'm tripping

I can only find this

oh, yeah - you should delete your kubernetes resource stack before deleting the cluster

The

Cluster failed to report its version number

means it can't access the k8s cluster

well, i really think you should have force delete resources option

@gorgeous-egg-16927 i know. its deleted, lol

We have

pulumi state delete

for that purpose: https://pulumi.io/reference/cli/pulumi_state_delete.html

If you want to just remove stuff from your stack

let me see if we have an issue for that situation. it seems like people will run into that if they're managing the cluster in a separate stack

ok, can you add

--all

parameter? 🙂

i dont think it makes any sense to manage kubernetes and applications on kubernetes in the same stack

its just unmanageable

what if I want new namespace for each branch?

what if i have 50 applications each with 20 namespaces? 🙂

Pulumi isn’t opinionated at all, you can totally do that if you want