Terraform: How to Upgrade Provider without Changing Terraform Version

Question

when I try to run terraform apply, I see Error:

aws_glue_catalog_database.test: Provider doesn't support resource: aws_glue_catalog_database

Looks like my provider is old because terraform version shows provider.aws v1.6.0 and in fact I can launch glue in another folder with the same terraform version but with a newer provider.aws v2.0.0

Terraform v0.11.10
+ provider.archive v1.0.0
+ provider.aws v1.6.0
+ provider.null v1.0.0
+ provider.template v1.0.0

I tried to upgrade provider but keep the terraform version v0.11.10 the same. to do that, I run terraform init -upgrade but see the below warning:

terraform init -upgrade
Initializing the backend...
Backend configuration changed!

Terraform has detected that the configuration     specified for the backend
has changed. Terraform will now check for existing state in the backends.


Do you want to migrate all workspaces to "s3"?
  Both the existing "s3" backend and the newly     configured "s3" backend
 support workspaces. When migrating between      backends, Terraform will copy
  all workspaces (with the same names). THIS WILL     OVERWRITE any conflicting
  states in the destination.

  Terraform initialization doesn't currently migrate     only select workspaces.
  If you want to migrate a select number of     workspaces, you must manually
  pull and push those states.

  If you answer "yes", Terraform will migrate all     states. If you answer
  "no", Terraform will abort.

I decided to say "no" because the above warning scared me.

I do have a backend "s3" resource to remote store the status in s3 and have several workspaces. I do not understand why and how the backend and workspace will be changed if I upgrade provider using above command; whether it will break my system.

Does anybody know whether it is ok for me to say yes without messing things? or which terraform cmd I should run to upgrade provider without change terraform version? Thanks.

hhami hhami · Accepted Answer · 2019-03-08T16:31:29

You're doing a major release upgrade so there is always risk I'm afraid.

Here are some links that could help you to highlight risks (if you haven't seen them already):

Before doing anything, I would recommend backing up the state files that are stored on the remote S3 bucket first. If your infrastructure is fairly static (or you have a mechanism to make it static), you could always put the old backups of the state back into S3 if it goes horribly wrong without causing an issue as no applies would have occurred during your upgrade process.

I don't know your setup but ideally you would be doing this in a development environment first which should hopefully ease your nerves about changing state files.

Terraform: How to Upgrade Provider without Changing Terraform Version

1 Answers