RDS Database Stuck in "Creating" Status for Days after Aurora Upgrade

Question

We recently attempted the AWS Aurora 2 (MySQL 5.7) -> Aurora 3 (MySQL 8) upgrade process described here: https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/AuroraMySQL.Updates.MajorVersionUpgrade.html#AuroraMySQL.Updates.MajorVersionUpgrade.2to3

It was successful for two of our dev environments. When we attempted to restore the Production Aurora 2.0/MySQL 5.7 snapshot in the MySQL 8.0 engine however, the creation process became stuck in the "Creating" state.

We attempted a second restore of the snapshot with the same results. We now have two Aurora 3.0 clusters in the "Creating" status, and they have remained this way for 3 days. We cannot delete, disable or modify these instances in any way.

Is there anything we can do to remove or recover these trapped instances?
Why might this bad restore happen? Our production DB isn't large - <1GB in size. It's only marginally older than our other environments.

Any help here would be much appreciated!

Callum Strados · Accepted Answer

The RDS team were able to resolve this for us:

> For an Aurora MySQL version 2 cluster that uses backtrack, currently you can't use the snapshot restore method to upgrade to Aurora MySQL version 3. This limitation applies to all clusters that use backtrack, regardless of whether the backtrack setting is turned on. In this case, perform a logical dump and restore by using a tool such as the mysqldump command.

AWS-User-9657330 · Answer

Just encountered the same thing, would be nice if the AWS API rejected restoring from a snapshot that had backtrack enabled. Now we've got a cluster stuck in creating.

Can someone at AWS please delete the cluster `qwertee-issue-3434` that's stuck on our account?

RDS Database Stuck in "Creating" Status for Days after Aurora Upgrade

Relevant content