Skip to content

Conversation

@yjhjstz
Copy link
Member

@yjhjstz yjhjstz commented Jan 5, 2026

Increase MAX_QUEUE_SIZE from 64 to 128 to allow a larger unacknowledged packet window in UDP interconnect, reducing stalls under high-latency or bursty workloads.

Fix calculation of per-queue capacity in SetupUDPIFCInterconnect by properly converting gp_interconnect_mem_size from MB to bytes. Without this conversion, the derived send queue depth could be severely underestimated, leading to suboptimal throughput.

Fixes #ISSUE_Number

What does this PR do?

Type of Change

  • Bug fix (non-breaking change)
  • New feature (non-breaking change)
  • Breaking change (fix or feature with breaking changes)
  • Documentation update

Breaking Changes

Test Plan

  • Unit tests added/updated
  • Integration tests added/updated
  • Passed make installcheck
  • Passed make -C src/test installcheck-cbdb-parallel

Impact

Performance:

User-facing changes:

Dependencies:

Checklist

Additional Context

CI Skip Instructions


Copilot AI review requested due to automatic review settings January 5, 2026 18:02
Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR improves the UDP interconnect performance by increasing the maximum queue size and fixing a critical unit conversion bug in memory calculation.

  • Doubles MAX_QUEUE_SIZE from 64 to 128 to support larger unacknowledged packet windows
  • Fixes memory size calculation by properly converting gp_interconnect_mem_size from MB to bytes (multiplying by 1024 * 1024)

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

…dling

Increase MAX_QUEUE_SIZE from 64 to 128 to allow a larger unacknowledged
packet window in UDP interconnect, reducing stalls under high-latency
or bursty workloads.

Fix calculation of per-queue capacity in SetupUDPIFCInterconnect by
properly converting gp_interconnect_mem_size from MB to bytes. Without
this conversion, the derived send queue depth could be severely
underestimated, leading to suboptimal throughput.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants