Fix(rdb-user- privileges): add check user and retry process

[Monitob]

No description provided.

[bLuka]

Hello,

I wanted to understand properly, and looked at your changes. I don't see any change from the place where the panic from #1054 happened. Do you mean the issue happened because of a failure during resource creation?

Thank you very much in advance for your answers!

[Monitob]

Hello,

I wanted to understand properly and look at your changes. I don't see any change from the place where the panic from #1054 happened. Do you mean the issue happened because of a failure during resource creation?

Thank you very much in advance for your answers!

Yes, I've noticed that the error from the API was not the right one.

Important points to understand the bug.

• The database name in the query was the instance name. (failure during the resource creation and the tfstate was not good)
• The code HTTP was 409 with not a clear message. Only state unavailable
• The privilege is a resource related to a user and a database. (this both resource doesn’t have an ID to work with)

Conclusion:

• The 409 means the user does not exist or the instance is really doing something so wait
• We have to check if the user really exist, otherwise, we skip
• To remove a user we need to put their privileges to none.
• The user and privilege could happen really fast so we check every time and it could be async.

The terraform provider allows us to manage the state of the resources and there are some cases that the state checks are not made.

Thank you for your @bLuka question and allow me to identify this. Please feel free to create a new issue for further fixes or contact me by our community slack.