bneradt commented on PR #13409:
URL: https://github.com/apache/trafficserver/pull/13409#issuecomment-5417440140

   Thanks @masaori335 for forwarding the concurrency and test audit. I amended 
the commit to address the actionable items:
   
   - RefCountCache now owns its public read/write locking, guarded state is 
annotated, and the unlocked erase/callback paths no longer race or 
self-deadlock.
   - HostDB RPCs return a server error before cache initialization, and 
successful clears write an audit note.
   - The AuTest now checks the exact empty-cache metric, a delayed in-flight 
transaction across a clear, repopulation without restart, extra-argument 
rejection, and the audit log. I also corrected the final process-liveness 
assertions.
   - The docs now clarify that host-file entries are not cleared, and the dead 
copy helper was removed.
   
   I considered moving the clear loop to ET_TASK. Running the same loop there 
would move the work but would not shorten any partition writer-lock hold, so I 
kept the operation synchronous and bounded one partition at a time. A 
chunked/asynchronous clear would be a reasonable follow-up if production 
measurements show material control-plane stalls.
   
   Validation on the rebased commit in the asfats5 container:
   
   - Full Debug CMake build and install: passed
   - test_RefCountCache: passed
   - hostdb_clear AuTest, including all 8 runs: passed
   
   The branch is rebased on current master at 052d4a4984.


-- 
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