Can't Modify RDS Postgres "R6i" Stand-Alone instance to Multi-AZ, Fails

0

I am working on some terraform to launch RDS instances. As part of this project I need to import data from another database. The recommended setup from AWS to have a FAST pg_restore is to launch the RDS instance in Standalone mode for the import and then modify to Mutli-AZ there after.

I have done this process before in my career and the methodology does work quite well. With this project we have chosen the R6i instance types as the match quite well with where we are migrating from. How ever I feel that we have found a bug in the AWS provisioning system. If you launch an RDS Postgres db.r6i.24xlarge instance type in the Stand-Alone Configuration and the modify it to be Multi-AZ you get the following error in the Web Console RDS - Log & Events Tab.

Instance type db.r6i.24xlarge in availability zone us-east-1c is unsupported; putting database instance into available

We know that to be false since the Main DB launched just fine.

I have tested using the AWS Web Console instead of the Terraform code and I can reproduce this. I have also found via the Web Console that you can get a Multi-AZ DB, but only if you start with that configuration. So the steps to reproduce:

  1. Launch a, RDS Postgres db.r6i.24xlarge instance in Stand-Alone.
  2. After Completion; Modify the instance to be Multi-AZ. The process with take a little over 30 min, but then you will get the above error.

My question: Is this a Bug?, and how can we get past it?

Update: I also Attempted the same tests in the Default VPC. In the Default both Starting as a Mutli-AZ AND Starting as a Stand-Alone and Modify failed. :-(

1 Answer
0
Accepted Answer

This is not a bug. This error is as a result of not having enough of the chosen instance class in the chosen availability zone at the time of request [1]. When the issue occurs, our internal team works at the backend to increase the capacity so that you will be able to create the instance later on.

Having said that, below could be the potential workarounds.

  1. Wait a while and then try again after a few minutes; capacity can shift frequently.
  2. Create instances in a different availability zone. (This applies when you creating a new instance)
  3. Try to scale up/down to a different instance class.
AWS
D-Rao
answered 2 years ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions