lizhimins commented on PR #4572:
URL: 
https://github.com/apache/rocketmq-dashboard/pull/4572#issuecomment-5761157602

   Thanks for the precise root-cause analysis — it matches ours exactly, and 
independently arriving at the same explanation is a good sign the flake was 
understood rather than papered over.
   
   `rmq_studio_session.last_seen_at` is a second-precision `DATETIME`, MySQL 
rounds a sub-second value up on insert, and the `idleSeconds >= 60` bound then 
occasionally reads back as 59. That is the right diagnosis.
   
   We are closing this as already covered on the development trunk. 
`AuthServiceSessionOverviewIntegrationTest` on `rocketmq-studio` now derives 
`now` with `LocalDateTime.now(ZoneOffset.UTC).withNano(0)` at both call sites, 
and carries a comment recording the same root cause ("the session columns are 
`datetime` without fractional precision and MySQL ROUNDS on insert"). For a 
`LocalDateTime`, `withNano(0)` and your `truncatedTo(ChronoUnit.SECONDS)` are 
equivalent, so the flake you set out to fix is gone. That change came in with 
#4606, the hosted-conversation rewrite that rebuilt the branch.
   
   Two practical notes:
   
   - All Studio PRs now merge onto the `rocketmq-studio` branch rather than 
`master`, and this PR is based on `master`. Retargeting it would only produce a 
conflict on the two lines that already carry the equivalent fix, with no 
behavioural delta.
   - There is no second site to sweep: we checked, and no other test writes a 
raw `LocalDateTime.now(ZoneOffset.UTC)` into a session column and then asserts 
on rounded bounds.
   
   If you still see this failure against the current `rocketmq-studio` HEAD, 
please reopen with the failing output — that would mean the rounding shows up 
somewhere we have not found, and we would want to know.
   


-- 
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