Powerdns record with count and hyphen in name fails on refresh

[alkersan]

provider fails to retrieve record information if name contains hyphens.

resource "powerdns_record" "hyphened" {
    count = "2"
    zone = "example.org"
    name = "redis-${count.index}.example.org"
    type = "A"
    ttl = 60
    records = [ "10.20.30.${count.index}" ]
}

Reason: poor choice of recordId separator character here.
Powerdns don't include unique identifier of rrsets, thus a tuple (name, type) was used to correlate terraform's resourceId with pdns record. But I totally missed the factor of allowed dns hostname characters and 've put allowed hyphen as separator between name and type, thus making impossible reverse parsing of resourceId like redis-1.example.com-A

While fix seems simple - to choose different separator (*$#&) that's not it allowed set, i understand this is a complete FUBAR. Change of internal resource representation will make all existing records effectively invalid after update. Is there some internal migration mechanism in terraform that can be applied? But if not - I don't think this provider attracted lots of users, so impact shouldn't be huge.

[jen20]

Fixed by #5229.

[jasonl]

For anyone who finds this, and needs to manually update their state file:

1. Open terraform.tfstate in a text editor
2. Find all the powerdns_record entries in it.
3. In each one, change both id keys from "your-domain.something.com-A" to "your-domain.something.com:::A".
4. Incremental the serial number at the top of terraform.tfstate so that Terraform doesn't think you've got a state file conflict.
5. Run terraform plan. It should show that no changes are required in your infrastructure.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.