From 39b04d3829739a62022f0999f1bd81e656cfd874 Mon Sep 17 00:00:00 2001 From: nevaforget Date: Fri, 24 Apr 2026 09:05:46 +0200 Subject: [PATCH] ci(build): drop unverified OOM throttle, run cargo with -j2 The previous CARGO_BUILD_JOBS=1 + nice + ionice throttling rested on an assumed OOM diagnosis for run 86 (2026-04-23) that was never confirmed via dmesg/journalctl. Build times tripled for no verified reason. Remove the throttle; if a real OOM is captured in the future, re-add limits with actual evidence. --- .gitea/workflows/build-and-publish.yaml | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/.gitea/workflows/build-and-publish.yaml b/.gitea/workflows/build-and-publish.yaml index 7533997..1db2cb7 100644 --- a/.gitea/workflows/build-and-publish.yaml +++ b/.gitea/workflows/build-and-publish.yaml @@ -53,14 +53,14 @@ jobs: sudo pacman -S --needed --noconfirm $MAKEDEPS fi - # Resource-constrained build. The act_runner shares CPU/RAM/I/O - # with the Gitea host (network-host mode). A parallel Rust build - # has OOM-killed the host: run 86 on 2026-04-23 stopped mid-compile - # with no error, taking gitea HTTPS down ~11 min. Force single-job - # compile and low CPU/IO priority so the host stays responsive. - export CARGO_BUILD_JOBS=1 - export MAKEFLAGS="-j1" - nice -n 19 ionice -c 3 makepkg -sfd --noconfirm + # Parallel build with two cargo jobs. Previous single-threaded + # throttling was based on an unverified OOM assumption (run 86 + # on 2026-04-23 stalled; no dmesg/journalctl evidence of + # OOMKiller was ever captured). If a real OOM happens, capture + # `dmesg | grep -i "killed process"` first before re-throttling. + export CARGO_BUILD_JOBS=2 + export MAKEFLAGS="-j2" + makepkg -sfd --noconfirm # makepkg can emit multiple artifacts per build (main + -debug # split package). Upload each. Arch filename convention: