Problem on json unmarshalling to go struct (MessageResource.message_list.details of type map[string]interface {})

[kruftik]

Nutanix Cluster Information

Nutanix Prism Central 5.17.0.3

Terraform Version

Terraform v0.13.5
+ provider registry.terraform.io/terraform-providers/nutanix v1.1.0

Initializing provider plugins...
- Using previously-installed terraform-providers/nutanix v1.1.0

Terraform has been successfully initialized!

Affected Resource(s)

It seems that the issue affects any Nutanix resource since it occurs on read phase of terraform plan / terraform apply commands. The cluster and subnet resources are 100% affected.

Terraform Configuration Files

data "nutanix_cluster" "cluster" {
  cluster_id = var.nutanix_cluster_uuid
}

Debug Output

can be provided upon request

Expected Behavior

terraform successfully fetch the actual cloud state of the resources

Actual Behavior

The following error occurs: Error: json: cannot unmarshal string into Go struct field MessageResource.message_list.details of type map[string]interface {}

Steps to Reproduce

run a command that causes terraform to request current state of the cloud resources (terraform apply / terraform plan)

[yannickstruyf3]

Hi @kruftik,

Is the data source for nutanix_cluster always failing (in 100 percent of the runs)?
Are you testing with a local user (admin?) or with an AD/LDAP user?

[kruftik]

Hi @yannickstruyf3,

Is the data source for nutanix_cluster always failing (in 100 percent of the runs)?

no. definitely, the errors happens not at each and every request to the API, I'd say it fails on 60% of the tries or so

Are you testing with a local user (admin?) or with an AD/LDAP user?

the account used is an AD-based one

[yannickstruyf3]

@kruftik Thanks for the fast reply!

For testing purposes, would it be possible to check if you have the same behaviour with a local (non-ad) account?
I would also recommend to create a Nutanix support ticket to validate the API failures on Prism Central side.

[kruftik]

@yannickstruyf3,

we tried to reproduce the issue using local nutanix account and didn't succeed.

moreover, we've identified the root cause of the problem - some of AD-controllers were not reachable from the Nutanix infrastructure. As soon as all the controllers became reachable, the problem has gone.

thanks a lot for a tip!