mirror of
https://github.com/kvcache-ai/Mooncake.git
synced 2026-11-05 01:17:55 +00:00
[PR #388] [MERGED] make backlog size of handshake listen configurable #1532
Labels
No labels
ascend
auto-closed
bug
bug
CI/Build
Common
documentation
enhancement
feature request
good first issue
help wanted
high priority
MetaX
nightly-failure
pull-request
PyTorch Backend
question
RFC
Roadmap
stale
Store
Tests
Transfer Engine
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
kvcache-ai/Mooncake#1532
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
📋 Pull Request Information
Original PR: https://github.com/kvcache-ai/Mooncake/pull/388
Author: @jiafuzha
Created: 5/22/2025
Status: ✅ Merged
Merged: 5/29/2025
Merged by: @doujiang24
Base:
main← Head:config_handshake_listen_backlog📝 Commits (7)
45502b1make backlog size of handshake listen configurable90f3cceUpdate mooncake-transfer-engine/src/config.cppca0b4dfRevert "Update mooncake-transfer-engine/src/config.cpp"
f629396Merge branch 'config_handshake_listen_backlog'ece6661make backlog size of handshake listen configurable35f807aMerge branch 'main' into config_handshake_listen_backlogd8fdab1Update mooncake-transfer-engine/include/config.h📊 Changes
6 files changed (+21 additions, -4 deletions)
View changed files
📝
doc/en/transfer-engine.md(+1 -0)📝
doc/zh/transfer-engine.md(+1 -1)📝
docs/source/design/transfer-engine.md(+1 -1)📝
mooncake-transfer-engine/include/config.h(+1 -0)📝
mooncake-transfer-engine/src/config.cpp(+11 -0)📝
mooncake-transfer-engine/src/transfer_metadata_plugin.cpp(+6 -2)📄 Description
When transfer engine configured with more cards, e.g., 8, via auto discovery, we get "connection reset by peer" error during handshaking. From dmesg, it shows "request_sock_TCP: Possible SYN flooding on port 0.0.0.0:16023. Sending cookies.". It means we need to increase backlog size of the listen API which is hard-coded to 5. We should make it configurable.
🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.