This message was deleted.
# general
s
This message was deleted.
w
You can remove resources individually by URN;
pulumi state delete URN
It's a lot more surgical that editing the state file manually. Perhaps it helps resolve your problem?
s
maybe I have hundred resources that depends on one resource that pulumi has not managed to cleanly delete
I mean, I have done it, but it takes a lot of time, and if this happens on a production situation it is going to be quite stressing
I do not understand either why pulumi cannot proceed with certaing pending operations
for example, in this case the pending operation was to delete a cluster
and the cluster was still there
why cant pulumi check that the resource is there and try to proceed?
w
It's a good question. Maybe you want to open an issue on github? https://github.com/pulumi/pulumi There's similar repos for the provider specific frameworks as well (AWS, GCP, etc)
s
for instance, now I entered a seemly impossible situation to resolve where pulumi was not able to destroy the stack due too some CRD that would not delete
after several attempts I had to delete the cluster manually
but now, pulumi destroy would not find the cluster and get stucked
I think pulumi cli should have some tool to be able overcome this state
w
Sounds like a github issue to me.
s
but I don’t know how to report it, since it is quite indeterministic, probably due GKE not being super stable either
I would need to have a simple reproducible case
w
Not necessarily. To me, it sounds like you're asking for a feature request that is independent of your particular failure mode; "remove large numbers of failing resources that break a deploy", or something like that.
Your failuire mode is just a scenario where that feature is particularly useful.
I think they will be able to come up with comparable scenarios that they are familiar with.
s
ok
w
If they can't come up with something comparable they will have questions and you get to start the conversation towards clarity. Either way it's a win for you.