maytasm opened a new pull request, #19787:
URL: https://github.com/apache/druid/pull/19787
fix: remove synchronized lock contention on broker query planning hot path
### Description
Summary
- Replace synchronized with volatile on
BrokerViewOfCoordinatorConfig.getCurrentServersToIgnore() and
setDynamicConfig() to eliminate monitor contention on the broker query planning
hot path
- getCurrentServersToIgnore() is called per-segment during
groupSegmentsByServer() and computeResultLevelCachingEtag() — with hundreds of
data servers and hundreds of concurrent queries, threads pile up on the monitor
causing a convoy effect
- synchronized was unnecessary: each read touches only one ImmutableSet
reference (selected by CloneQueryMode), and volatile provides sufficient
visibility for immutable reference swaps. This is safe because:
- ImmutableSet.copyOf() creates fully-constructed immutable objects
- volatile guarantees visibility of the reference swap
- Readers may briefly see a stale config during setDynamicConfig(), but
that's acceptable — setDynamicConfig() is called infrequently (config updates),
and a query seeing the old config for one cycle is harmless
Evidence
- Flamegraph from production showed 459 concurrent queries with 77.8% of
wall time BLOCKED on getCurrentServersToIgnore monitor
- Thread dump showed 33 threads blocked on the same monitor, 105
ForkJoinPool threads idle (starved)
- Queries that run in ~3s without load degraded to ~20s under load
- (Note: that our production queries hits around 300k segments per queries
amplifying this issue)
This PR has:
- [ ] been self-reviewed.
- [ ] using the [concurrency
checklist](https://github.com/apache/druid/blob/master/dev/code-review/concurrency.md)
(Remove this item if the PR doesn't have any relation to concurrency.)
- [ ] added documentation for new or modified features or behaviors.
- [ ] a release note entry in the PR description.
- [ ] added Javadocs for most classes and all non-trivial methods. Linked
related entities via Javadoc links.
- [ ] added or updated version, license, or notice information in
[licenses.yaml](https://github.com/apache/druid/blob/master/dev/license.md)
- [ ] added comments explaining the "why" and the intent of the code
wherever would not be obvious for an unfamiliar reader.
- [ ] added unit tests or modified existing tests to cover new code paths,
ensuring the threshold for [code
coverage](https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md)
is met.
- [ ] added integration tests.
- [ ] been tested in a test Druid cluster.
--
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]
