Failed to marshal plan to json: value is not known

[stjohnfinn]

Getting Failed to marshal plan to json: value is not known when I plan and the plan just completely stops there and it doesn't show N to create, N to change, N to delete, so it kind of makes the provider unusable. I don't think it's an issue with my code because I'm not doing anything weird or complex and I've had Claude starting at it for a couple hours. I think it's something with computed values....maybe the ID of these resources? but computed values don't break plans as far as I know. EC2 instance IDs are post-apply computed values and they don't break the plan.

Not really sure what's going on. I would love to use this provider but I can't really use it with this bug.

[stjohnfinn]

Seems to be related to imported resources. I tried importing a bunch of resources. When I run plan, it shows this message and does not show predicted changes. I was like "well it seems like other people are using this provider somehow", so I tried just importing one resource and then applying (without seeing the predicted changes) and it ended up fine. Now when I run plan, it shows predicted changes and I don't see the Failed to marshal message.

[stjohnfinn]

And yes I tried all sorts of flags, and different ways to view the plan, and applying to see if the plan that apply generates was usable, and none of it really worked.

[stjohnfinn]

Got it to work by "blindly" applying a single import and praying it wouldn't do any hard to the infrastructure. That worked fine and then the "unable to marshal" message went away. So I figured it had to do with imported resources, but I'm now starting to import more resources but I'm not getting that message, so maybe it was caused by having an empty state besides the imported resources and then trying to apply.

[stjohnfinn]

╷
│ Error: Client Error
│
│ with module.standard_proxy_hosts["xxxxx"].nginxproxymanager_proxy_host.host,
│ on ../modules/default_proxy_host/main.tf line 63, in resource "nginxproxymanager_proxy_host" "host":
│ 63: resource "nginxproxymanager_proxy_host" "host" {
│
│ Unable to update proxy host, got error: data failed to match schemas in oneOf(GetProxyHosts200ResponseInnerAccessList)
╵
Getting this too when applying just a change to a single proxy host which just updates the advanced config for a custom location. I don't know why.

[Sander0542]

Thanks for reporting this bug. This sound hard to replicate. Would you be able to write out a scenario that always fails?

[Sander0542]

╷ │ Error: Client Error │ │ with module.standard_proxy_hosts["xxxxx"].nginxproxymanager_proxy_host.host, │ on ../modules/default_proxy_host/main.tf line 63, in resource "nginxproxymanager_proxy_host" "host": │ 63: resource "nginxproxymanager_proxy_host" "host" { │ │ Unable to update proxy host, got error: data failed to match schemas in oneOf(GetProxyHosts200ResponseInnerAccessList) ╵ Getting this too when applying just a change to a single proxy host which just updates the advanced config for a custom location. I don't know why.

Which version of nginx proxy manager are you using?