RDS Read Replica stuck in "incompatible-network" state - All operations failing

Question

Hi everyone,

I'm dealing with a PostgreSQL 17.4 read replica that's stuck in `incompatible-network` state after a subnet group issue. Looking for advice on how to resolve this.

## What happened:
The DB subnet group originally had invalid/deleted subnets. I've since fixed the subnet group and created a new working replica, but the original replica remains stuck.

## Current situation:
- **Region**: us-east-1
- **Engine**: PostgreSQL 17.4  
- **State**: incompatible-network
- **Created**: 2025-06-25

## Failed operations:
Every AWS CLI operation returns `InternalFailure` or `InvalidDBInstanceState`:
```bash
aws rds modify-db-instance  # (to disable deletion protection)
aws rds delete-db-instance
aws rds promote-read-replica
aws rds reboot-db-instance
aws rds create-db-snapshot
```

## What I've tried:
1. Fixed the original subnet group configuration
2. Created a new working replica as replacement
3. Set max allocated storage to prevent growth on primary

## Question:
Has anyone encountered this before? Is there a way to:
- Force the replica back to a healthy state, or
- Override deletion protection to clean up the stuck instance?

The replacement replica is working fine, so data loss isn't a concern for this specific instance.

Any insights would be appreciated!

re:Post Agent · Answer

To fix your RDS read replica that's stuck in an "incompatible-network" state, you should try using the `start-db-instance` command via the AWS CLI. This is the recommended approach for resolving this specific issue.

Here's how to fix it:

1. Open the AWS Management Console and navigate to the RDS section
2. Select the affected database from the Databases list
3. Note down the DB instance identifier, VPC ID, and subnet IDs from the Connectivity & Security tab
4. Use the AWS CLI to run the following command:
```
aws rds start-db-instance --db-instance-identifier your-instance-identifier
```

If successful, your database status should change from "incompatible-network" to "Available."

Be aware that if the database restarts, it might execute the last operation that was running before it entered the incompatible-network state. This could potentially cause the instance to return to the incompatible-network state if that operation was the cause of the issue.

If this approach doesn't work, you may need to contact AWS Support for further assistance, as they might need to help you override the deletion protection or resolve the underlying network configuration issue that's preventing normal operations.  
**Sources**  
[Troubleshooting for Amazon RDS - Amazon Relational Database Service](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_Troubleshooting.html)  
[Resolving identifiable vacuum blockers in RDS for PostgreSQL - Amazon Relational Database Service](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Appendix.PostgreSQL.CommonDBATasks.Autovacuum_Monitoring.Resolving_Identifiableblockers.html)  
[Resolving identifiable vacuum blockers in Aurora PostgreSQL - Amazon Aurora](https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Appendix.PostgreSQL.CommonDBATasks.Autovacuum_Monitoring.Resolving_Identifiableblockers.html)

Manvitha Potluri · Answer

You need to open a support ticket with AWS and request one of the following:

a) Force delete the stuck read replica from backend (regardless of its state or deletion protection).
b )Remove the “incompatible-network” status manually from their side to allow operations again.

OR
**What You Can Do**
Check if deletion protection is ON:

If yes, and modify-db-instance fails with InvalidDBInstanceState, you cannot disable it via CLI.

This confirms the instance is essentially “locked” in an unrecoverable state from the customer side.

Promote the read replica?

Not possible: AWS won’t allow promoting a replica in incompatible-network.

Force delete with CLI?

No. even with --skip-final-snapshot, it fails due to instance state.

Start/Stop workaround?

Not allowed for read replicas, confirmed by your error.

Riku_Kobayashi · Answer

Hello.

Have you tried the troubleshooting steps in the following document?  
The fix is ​​described in the AWSSupport-ValidateRdsNetworkConfiguration runbook.  
https://repost.aws/knowledge-center/rds-incompatible-network

RDS Read Replica stuck in "incompatible-network" state - All operations failing

What happened:

Current situation:

Failed operations:

What I've tried:

Question:

Add your answer

Relevant content