terraform-provider-redshift
terraform-provider-redshift copied to clipboard
warnings from recent versions of terraform
When I look at the log, I see a large number of warnings about non-computed attribute
like
2021-12-21T14:22:02.999-0100 [WARN] Provider "registry.terraform.io/brainly/redshift" produced an in
valid plan for redshift_user.xyz, but we are tolerating it because it is using the legacy plugin
SDK.
The following problems may be the cause of any confusing errors from downstream operations:
- .superuser: planned value cty.False for a non-computed attribute
- .syslog_access: planned value cty.StringVal("RESTRICTED") for a non-computed attribute
- .valid_until: planned value cty.StringVal("infinity") for a non-computed attribute
- .connection_limit: planned value cty.NumberIntVal(-1) for a non-computed attribute
- .create_database: planned value cty.False for a non-computed attribute
I dont know if these are actual problems, but they make it less nice to read the logs :-)
Hi, could you share the provider and terraform version you are using? The warning message refers to "legacy plugin SDK" and according to docs is the SDK v1 while the provider code is using the SDK v2 from the very beginning.
@winglot hm, yes, that is strange... i'm using terraform-redshift-provider
v0.5.0 (also checked v0.5.1 and it still happens) and terraform v1.1.2 on darwin_amd64 ,go looking in log.out
, with settings
export TF_LOG=DEBUG
export TF_LOG_PATH=log.out
also happens with redshift_schema
:
2021-12-30T16:50:00.739-0600 [INFO] Terraform version: 1.1.2
2021-12-30T16:50:00.739-0600 [INFO] Go runtime version: go1.17.5
2021-12-30T16:50:00.739-0600 [INFO] CLI args: []string{"terraform", "plan"}
...
2021-12-30T16:50:02.949-0600 [INFO] provider.terraform-provider-redshift_v0.5.1: 2021/12/30 16:50:02 [DEBUG] Collected schema 'information_schema' privileges for group ...
...
2021-12-30T16:50:05.162-0600 [WARN] Provider "registry.terraform.io/brainly/redshift" produced an invalid plan for main.redshift_schema.asdf, bu
t we are tolerating it because it is using the legacy plugin SDK.
The following problems may be the cause of any confusing errors from downstream operations:
- .quota: planned value cty.NumberIntVal(0) for a non-computed attribute
- .external_schema[0].data_catalog_source[0].catalog_role_arns: planned value cty.ListValEmpty(cty.String) for a non-computed attribute
- .external_schema[0].data_catalog_source[0].region: planned value cty.StringVal("") for a non-computed attribute
~~Not sure if this helps, but I tried the migrator checker https://www.terraform.io/plugin/sdkv2/guides/v1-upgrade-guide#using-tf-sdk-migrator and got:~~ Edit: nevermind, this is all irrelevant.
terraform-provider-redshift % ~/go/bin/tf-sdk-migrator check
Checking Go runtime version ...
Go version 1.17.5: OK.
Checking whether provider uses Go modules...
Go modules in use: OK.
Checking version of github.com/hashicorp/terraform-plugin-sdk to determine if provider was already migrated...
Checking version of github.com/hashicorp/terraform used in provider...
This directory (/Users/mememe/src/terraform-provider-redshift) doesn't seem to be a Terraform provider.
Providers depend on github.com/hashicorp/terraform