terrymanu commented on issue #37514:
URL:
https://github.com/apache/shardingsphere/issues/37514#issuecomment-3691208739
Problem Understanding
- After an AWS MySQL failover (RDS/Aurora unspecified), writes via
ShardingSphere-Proxy hit “Running in read-only.” The error clears after some
time or after restarting Proxy. The user believes read/write config is correct.
Root Cause
- The open-source ShardingSphere does not auto-detect or switch
master/replica roles. After failover, the data source still points to the old
primary, which is now read-only. Writes routed there raise “read-only.” This is
expected behavior for static read/write
splitting, not a product bug.
Analysis
- Static read/write splitting does not refresh when instance roles change;
AWS failover sets the old primary to read-only, so existing connections keep
failing until the pool expires or Proxy is restarted.
- If you used AWS writer endpoints yet still saw this, you need to confirm
the endpoint’s post-failover resolution matches the addresses held by
ShardingSphere, and check Proxy logs during the failover window to see whether
it kept using the old address.
- Lacking version and configuration details, other misconfigurations can’t
be fully ruled out, but the observed pattern aligns with “static config not
updated after failover.”
Conclusion
- This is expected static read/write splitting behavior without automatic
master refresh; not a ShardingSphere defect.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
