Closed ann8ty closed 6 months ago
ok, my new theory is that this issue is actually caused by https://github.com/fivetran/terraform-provider-fivetran/issues/253 which between 1.1.10 and 1.1.16 i need to change SSHTunnel to SshTunnel. this worked for my dev and I am now testing on other environments:
yes, this resolved it.
Describe the bug The fivetran_destination output public_key is now blank. Can confirm in the fivetran web UI there is a public key for this destination (and the value matches what is currently stored in the secret version).
attempting to upgrade to v1.1.16 but cannot due to this blocker
Doc: https://registry.terraform.io/providers/fivetran/fivetran/latest/docs/resources/destination
To Reproduce Provide an example .tf configuration
snippet creates fivetran destination (it is already existing)
snippet saves public key to aws secrets manager for use with bastion automation this code previously worked, but with v1.1.16 the config.public_key is empty
run terraform output fivetran_public_key after plan, apply to see that public_key is empty
in v1.1.10 the code for config was indexed like this:
Expected behavior public_key should have value
Logs & Output Provide terraform CLI output on
plan
|apply
. If it is possible, please, refer to this guide to enable debug output.sanitized plan output
shows that terraform thinks the public key. has changed (it has, it is now blank, which is not good!)
then apply fails because... its blank
Plugin version: v1.1.16
Additional context
I wanted to see if this public_key was blank for a new redshift destination as well, and it is. So this problem exists both for an existing destination, and a new destination. You can ignore the bastion complaining about auth failures, the important bit is │ Error: putting Secrets Manager Secret value: InvalidRequestException: You must provide either SecretString or SecretBinary. indicating the public key is empty.