MySQL Aurora failover ends up on reader when using custom domain

[szhnet]

Describe the bug

#386 The problem still exists.

A jdbc url like this: jdbc:aws-wrapper:mysql://[custom-domain]:3306/db?characterEncoding=utf-8&clusterInstanceHostPattern=?.[clusterid].us-east-1.rds.amazonaws.com&wrapperLoggerLevel=ALL

In the log there is something like this
The driver is unable to track this opened connection because the instance endpoint is unknown: '[custom-domain]'

java.sql.SQLException: Running in read-only mode

Expected Behavior

The failover should end up on the writer node.

What plugins are used? What other connection properties were set?

Default plugins are used.

Current Behavior

The connection ends up on the reader instead of the writer.

Reproduction Steps

See description.

Possible Solution

No response

Additional Information/Context

No response

The AWS Advanced JDBC Driver version used

2.5.4

JDK version used

11

Operating System and version

Linux

[aaron-congo]

Hi @szhnet,

Thanks for reaching out and raising this issue. We'll take a look at this and keep you updated as we investigate. Thank you for your patience!

[karenc-bq]

Hi @szhnet, I am unable to reproduce the error you are seeing on v2.5.4. Would you be able to share the full driver logs with us?
We have recently merged some fixes related to custom domains. These fixes are available in this snapshot build: aws-advanced-jdbc-wrapper-2.5.5-20250206.155700-54.jar. While we are still investigating the root cause of this issue, would you be interested in giving this snapshot build a try to see if this resolves your issue?

[szhnet]

full log
app1.log

[karenc-bq]

Hi, @szhnet, thank you for the driver logs. It looks like driver failover wasn't triggered because no network errors were detected. Would you be able to provide more information about your custom endpoint and your test setup?

1. Does your custom endpoint point to a cluster endpoint or an instance endpoint?
2. How are you triggering cluster failover?

[szhnet]

1. In Route 53 add cname [custom-domain] -> xxx.cluster-[clusterid].us-east-1.rds.amazonaws.com

2. Use manual failover in the AWS Management Console.
   

3. My test program executes an update sql every 5 seconds. If the execution is successful, you can see something like "FooService work count=0" in the log. After count=3, when I use manual failover, you can see that "Running in read-only mode" starts to appear.

4. When I use "xxx.cluster-[clusterid].us-east-1.rds.amazonaws.com" in jdbc url, the test program is OK.

[longzhihun]

same here, when used Route53 domain cname to aurora endpoint, it doesn't work, do we have any workaround.

[longzhihun]

[karenc-bq]

Hi both, could you please try the snapshot build aws-advanced-jdbc-wrapper-2.5.5-20250206.155700-54.jar and set the verifyOpenedConnectionType=writer connection parameter to see if this resolves your issue?

[longzhihun]

@karenc-bq Hello, the link response 404

[sergiyvamz]

Hi @longzhihun

You can use any of the latest snapshot builds.
https://aws.oss.sonatype.org/content/repositories/snapshots/software/amazon/jdbc/aws-advanced-jdbc-wrapper/2.5.5-SNAPSHOT/aws-advanced-jdbc-wrapper-2.5.5-20250303.181434-75.jar

https://aws.oss.sonatype.org/content/repositories/snapshots/software/amazon/jdbc/aws-advanced-jdbc-wrapper/2.5.5-SNAPSHOT/

[szhnet]

Hello
I tried the aws-advanced-jdbc-wrapper-2.5.5-20250303.181434-75.jar and set verifyOpenedConnectionType=writer, the problem still exists.

app2.log