Use dedicated reactor for RaftReplService timers #854
Conversation
|
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #854 +/- ##
==========================================
- Coverage 56.51% 48.38% -8.13%
==========================================
Files 108 110 +2
Lines 10300 12818 +2518
Branches 1402 6152 +4750
==========================================
+ Hits 5821 6202 +381
+ Misses 3894 2532 -1362
- Partials 585 4084 +3499 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Pull request overview
This PR addresses potential deadlock issues by moving RaftReplService timers from global worker reactors to a dedicated reactor, preventing timer competition with high-priority IO tasks.
Changes:
- Replaced global timers with thread timers running on a dedicated "raft_repl_svc_timer" reactor
- Wrapped timer cancellation in
run_on_waitto ensure execution on the correct reactor fiber - Bumped version from 7.2.1 to 7.2.2
Reviewed changes
Copilot reviewed 2 out of 3 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| src/lib/replication/service/raft_repl_service.cpp | Refactored timer initialization to use dedicated reactor and thread timers instead of global timers |
| conanfile.py | Incremented version number to reflect the changes |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Global timers using `all_worker` regex can compete with default worker reactors, potentially jamming high-priority IO tasks and causing deadlocks. Signed-off-by: Jilong Kou <jkou@ebay.com>
4 participants
Global timers using
all_workerregex can compete with default worker reactors, potentially jamming high-priority IO tasks and causing deadlocks.Context: #847