nic-6443 opened a new pull request, #13553:
URL: https://github.com/apache/apisix/pull/13553

   ### Description
   
   Fixes #13454
   
   Sentinel-mode counterpart of #13516. The `redis-sentinel` policy goes 
through api7-lua-resty-redis-connector, which skips AUTH/SELECT on connections 
reused from the keepalive pool while pooling by the default `host:port` key — 
so two routes pointing at the same sentinel-managed master with different 
`redis_database` or credentials cross-contaminate exactly like the plain-redis 
case: limit counters land in the wrong database.
   
   This bumps the connector to 0.13.0 (api7/lua-resty-redis-connector#9), which 
derives the default pool name from everything the connection gets bound to at 
handshake time — resolved address, database, a credentials digest, and TLS 
options — while an explicit `connection_options.pool` still wins. Building the 
name from the resolved address keeps failover correct: a new master gets a new 
pool instead of reusing connections to the old one.
   
   The regression test mirrors the plain-redis one: two routes on the same 
sentinel master with `redis_database` 1 and 2 must keep their counters in their 
own databases. Without the bump it fails with both keys landing in db 1. The 
test helper now also flushes the sentinel master (port 6479) so counters don't 
survive CI re-runs.
   
   ### Checklist
   
   - [x] I have explained the need for this PR and the problem it solves
   - [x] I have explained the changes or the new features added to this PR
   - [x] I have added tests corresponding to this change
   - [ ] I have updated the documentation to reflect this change
   - [x] I have verified that this change is backward compatible


-- 
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]

Reply via email to