terraform-provider-snowflake: data_retention_time_in_days = 0 must be defined in snowflake_database from share

Terraform CLI and Provider Versions

Provider version: 0.85.0

Terraform Configuration

resource "snowflake_database" "from_share" {
  for_each                    = var.databases_from_share
  
  name                        = upper(each.key)
  # data_retention_time_in_days = 0
  from_share = {
    provider = upper(each.value.from_share.provider)
    share    = upper(each.value.from_share.share)
  }
}

Expected Behavior

After successfully creating the database from share I can add other resources and apply the changes.

Actual Behavior

After successfully creating the database from share I cannot add other resources and apply the changes. all successive applies try to set data_retention_time_in_days = 1: Terraform will perform the following actions:

  # snowflake_database.from_share["my_share"] will be updated in-place
  ~ resource "snowflake_database" "from_share" {
      ~ data_retention_time_in_days = 0 -> 1
        id                          = "MY_SHARE"
        name                        = "MY_SHARE"
        # (2 unchanged attributes hidden)
    }
    ...
snowflake_database.from_share["my_share"]: Modifying... [id=MY_SHARE]
╷
│ Error: error updating database data retention time on MY_SHARE err = 003001 (42501): SQL access control error:
│ Insufficient privileges to operate on database 'MY_SHARE'
│ 
│   with snowflake_database.from_share["my_share"],
│   on snowflake_shares.tf line 1, in resource "snowflake_database" "from_share":
│    1: resource "snowflake_database" "from_share" {
│ 

Adding data_retention_time_in_days = 0 to the database definition fixes this: Terraform doesn’t detect any changes in the database from share

Steps to Reproduce

  1. terraform apply without data_retention_time_in_days
  2. terraform apply another time: the second time terraform want to set data_retention_time_in_days = 0 -> 1

How much impact is this issue causing?

Medium

Logs

No response

Additional Information

No response

About this issue

  • Original URL
  • State: closed
  • Created 3 years ago
  • Comments: 22

Most upvoted comments

Hi @sfc-gh-jcieslak , I will close this issue. Thanks!

@sfc-gh-jcieslak this is the creation statement that is created by version 0.87 of the provider: CREATE DATABASE "MYSHARE" FROM SHARE "PROVIDER.SHARE"."SHARED_SCHEMA"; It should be: CREATE DATABASE "MYSHARE" FROM SHARE "PROVIDER"."SHARE"."SHARED_SCHEMA";

Ok, I didn’t analyze it in-depth, but this looks like it’s related to this issue - https://github.com/Snowflake-Labs/terraform-provider-snowflake/issues/2356. So, the issue may be happening, because there’s a default value set on data_retention_time_in_days (set to 1). When it’s not set in the config and remotely it’s set to something different than 1, Terraform tries to set it back to 1 (because that’s the default). I’m working on a change right now that will remove the default, so after a merge and release it should work as intended. As you already pointed out the workaround is to set it explicitly to 0.