fix: write each SPOP frame with a single socket write - #1
Merged
Conversation
Signed-off-by: Oleksander Piskun <oleksandr2088@icloud.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The library wrote each SPOP frame with three separate write() calls (length prefix, frame header, payload). asyncio sends every write() immediately with TCP_NODELAY, so frames regularly left as 2-3 TCP segments. HAProxy 3.2's rewritten SPOP mux has a timing-sensitive bug with an AGENT-HELLO split across segments: it resets the connection instead of waiting for the remaining bytes, so SPOE transactions fail intermittently once all connection retries are lost. Verified: roughly 25-40% of cold connection attempts failed on HAProxy 3.2.0/3.2.22/3.3/3.4, zero on 3.1.2 (whose legacy SPOE applet reassembled partial frames fine).
Emitting the whole frame in one write removes the partial-frame window entirely and makes HaRP on HAProxy 3.2 fully stable: 15/15 clean cold starts and zero handshake resets, against ~30% failed cold starts before. One syscall instead of three is also simply better.
Found while investigating the CI failure on nextcloud/HaRP#116.