Serialize libpq access to the shared replication connection#102
Merged
Conversation
Contributor
📊 Benchmark ResultsCurrent run is the minimum over 3 passes, compared against the base branch (
Summary: ➡️ 9 neutral · 🔴 2 slower · ⚪ 2 ignored (sub-μs) Thresholds: <1μs ignore · 1–20μs 15% · 20–50μs 10% · ≥50μs 5%. Measured on a shared CI runner — treat small deltas as noise. Informational only; this check never fails the build. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
The receive loop and the flush worker share one
PGconn: the main thread callsPQgetCopyData/PQconsumeInput, the worker sends standby feedback viaPQputCopyData/PQflush. libpq forbids using one connection from two threads: internal buffers are reallocated with no locking, and a blocked send even reads input from the writer thread (pqSendSomecallspqReadDataon EWOULDBLOCK). Under load this race wedged the connection:PQflushstuck inselect()forever and the process hung instead of failing fast.Solution
std.Io.MutextoReplicationProtocol; every libpq call after startup takes it: the receive path, the feedback path, and thePQerrorMessagereads on error paths.poll()wait stays outside the lock, so feedback never queues behind an idle wait. Contention is negligible: the worker locks once per flush interval to send 34 bytes.lockUncancelableon purpose: the critical sections are microseconds, and a cancelation point inside a libpq call would let shutdown interrupt the final LSN commit.PGconn(buffer reallocs, error state), so shared locking would recreate the race.Simpler alternative to #92: threading stays as is, only the connection access is serialized.