rajvarun77 commented on PR #3323:
URL: https://github.com/apache/brpc/pull/3323#issuecomment-4597467618

   ### The failing `clang-unittest-asan` job is a pre-existing test race — not 
caused by this change
   
   This PR only **installs** `redis-server` + `mysql-server`. That install step 
passes in every job it was added to (`clang-unittest` ✅, `clang-unittest-asan` 
✅). The single red job fails inside `brpc_redis_unittest`, on the 5 tests that 
talk to a redis the **test itself** forks on port `6479`:
   
   ```
   brpc_redis_unittest.cpp:173: Failure
   Value of: cntl.Failed()   Actual: true   Expected: false
   [E111]Fail to connect ...addr=0.0.0.0:6479...: Connection refused
   [  FAILED  ] RedisTest.sanity / keys_with_spaces / incr_and_decr / 
by_components / auth
   ```
   
   **Why:** `RunRedisServer()` in `test/brpc_redis_unittest.cpp` forks 
`redis-server --port 6479` and then waits a fixed `usleep(50000)` (50 ms) 
before the first client test connects. On the ASAN runner redis only becomes 
ready ~82 ms after the fork (same job log: fork ≈ `21:00:06.93` → `Ready to 
accept connections` `21:00:07.013`), so the first 5 client tests hit 
*Connection refused*.
   
   Before this change, `which redis-server` failed, so those 5 tests were 
**skipped** (auto-green). Installing redis simply **un-skipped** them and 
exposed the latent 50 ms startup race. The race lives in 
`test/brpc_redis_unittest.cpp`, independent of this workflow change.
   
   ![evidence: PR asan failure is a pre-existing redis test 
race](https://raw.githubusercontent.com/rajvarun77/brpc/ci-debug-evidence/ci-debug/redis_race_evidence.png)
   


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to