[PR #2333] [CLOSED] do not retry NXDOMAIN over TCP when using `try_tcp_on_error` #2964

New issue

Closed

opened 2026-03-16 11:17:42 +03:00 by kerem · 0 comments

kerem commented

2026-03-16 11:17:42 +03:00

Owner

📋 Pull Request Information

Original PR: https://github.com/hickory-dns/hickory-dns/pull/2333
Author: @lrouquette
Created: 7/30/2024
Status: ❌ Closed

Base: main ← Head: no-retry-nxdomain

📝 Commits (5)

c77ae08 do not retry NXDOMAIN over TCP when using try_tcp_on_error
297312c simplify match case
5b29501 fmt
29c73a5 Merge branch 'main' into no-retry-nxdomain
07e5b30 only if trusted

📊 Changes

2 files changed (+117 additions, -0 deletions)

View changed files

📝 crates/resolver/src/name_server/name_server_pool.rs (+14 -0)
📝 tests/integration-tests/tests/name_server_pool_tests.rs (+103 -0)

📄 Description

All the background info is in #2228. At scale, retrying over TCP is adding quite a bit of latency.

The issue suggests changing to if e.is_no_connections() || (opts.try_tcp_on_error && e.is_io()) but I was wondering if it was too broad. Happy to put this behind a config flag to keep the existing behavior if needed.

Fixes: #2228

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/hickory-dns/hickory-dns/pull/2333 **Author:** [@lrouquette](https://github.com/lrouquette) **Created:** 7/30/2024 **Status:** ❌ Closed **Base:** `main` ← **Head:** `no-retry-nxdomain` --- ### 📝 Commits (5) - [`c77ae08`](https://github.com/hickory-dns/hickory-dns/commit/c77ae08ae38172e5e9556ff171e1da6cbcba9c7c) do not retry NXDOMAIN over TCP when using try_tcp_on_error - [`297312c`](https://github.com/hickory-dns/hickory-dns/commit/297312ceaad0ad41afc60a2398eece2aa7a2bd1a) simplify match case - [`5b29501`](https://github.com/hickory-dns/hickory-dns/commit/5b2950100f498b9e035f08d31a98efe0da1922ec) fmt - [`29c73a5`](https://github.com/hickory-dns/hickory-dns/commit/29c73a56d03410152a20dc6c10f8598f61a87065) Merge branch 'main' into no-retry-nxdomain - [`07e5b30`](https://github.com/hickory-dns/hickory-dns/commit/07e5b30f84b405d7e75f51cf88744e45517d92f9) only if trusted ### 📊 Changes **2 files changed** (+117 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `crates/resolver/src/name_server/name_server_pool.rs` (+14 -0) 📝 `tests/integration-tests/tests/name_server_pool_tests.rs` (+103 -0) </details> ### 📄 Description All the background info is in #2228. At scale, retrying over TCP is adding quite a bit of latency. The issue suggests changing to `if e.is_no_connections() || (opts.try_tcp_on_error && e.is_io())` but I was wondering if it was too broad. Happy to put this behind a config flag to keep the existing behavior if needed. Fixes: #2228 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>