Why does an AWS RDS subnet group require multiple availability zones?

Question

When creating a new RDS MS SQL instance it requires a subnet group that spans multiple availability zones. I understand that this is different to Multi A-Z replication and that my instance will exist within a single availability region. So what is this actually doing when I'm adding this subnet group spanning different zones?

Witmer

Answer 1

RDS uses the additional availability zone for storing backup data and logs for redundancy, the data for your instance is stored within the other availability zone.

More info can be found here:

https://aws.amazon.com/rds/faqs/

Q: Why does additional backup storage cost more than allocated DB Instance storage?

The storage provisioned to your DB Instance for your primary data is located within a single Availability Zone. When your database is backed up, the backup data (including transactions logs) is geo-redundantly replicated across multiple Availability Zones to provide even greater levels of data durability. The price for backup storage beyond your free allocation reflects this extra replication that occurs to maximize the durability of your critical backups.

Cumbrous

Please justify your answer with a specific pull-quote from the linked page. Outside the context of multi-AZ, I don't see that what you are saying can be found at that link. – Fiji

@Michael-sqlbot I've updated with the quote, it's quite easy to find I'm not sure why you had any trouble finding it yourself. – Cumbrous

Makes sense, I guess that this also explains why a Multi A-Z deployment requires 3 zones (1 for each database and another for the logs and back-ups). – Witmer

I couldn't find it, because this seems like a non sequitur -- that quote appear doesn't have anything to do with subnet groups. The infrastructure on back side that's storing this data has no obvious relationship at all to which subnets in which availability zones you select for RDS to use for deploying instances into. – Fiji

Answer 2

There are separate plans for [Multi-AZ deployment (one standby)] and Multi-AZ deployment (two readable standbys). so it can't be that the other AZ will be used as standby since you're deploying a single AZ instance.

According to FAQs, RDS Backups are stored in S3. [Amazon RDS FAQs]

Q: Where are my automated backups and DB snapshots stored and how do I manage their retention? A: Amazon RDS DB snapshots and automated backups are stored in S3.

as it turns out, there's no specific reason [again, from the same FAQs page]--

Then why does it need you to define other AZ? For Multi-AZ deployments, defining a subnet for all Availability Zones in a Region will allow Amazon RDS to create a new standby in another Availability Zone should the need arise. You need to do this even for Single-AZ deployments, just in case you want to convert them to Multi-AZ deployments at some point.

The second subnet is just sitting there doing nothing waiting for you to upgrade to Multi-AZ. This is just how it is designed and implemented, to ease the conversion later if needed. Probably to make migration to Multi-AZ as simple as possible. [aws re:post]

In case of failure, will AWS automatically transition the RDS to the second subnet? -- No, it will only transition to the second subnet if Multi-AZ is enabled.

Doxia