awkawbandClaude Sonnet 5
b1fe86e9d5
Root-caused a "logo upload hangs, connection refused in browser console" report: the qemu-vm module's defaults (1 core, 1024MB) are too tight for this stack (MySQL + Redis + PHP-FPM + chromium for PDF generation, each PDF request spawning a full multi-process browser instance) under real concurrent load, and this VM boots with panic_on_oom - so exhausting memory doesn't degrade gracefully, the whole VM goes down and every connection starts refusing. Reproduced directly: at the 1024MB default, even 2 concurrent PDF generation requests triggered `redis-server invoked oom-killer` and a kernel panic. Bumped to 2 cores / 2560MB, which comfortably survives that same load (~1.4GB still free afterward) - verified via the exact concurrent-request scenario that crashed the 1024MB config, not just absence of errors. virtualisation.cores/memorySize aren't options on a plain nixosSystem in this nixpkgs (only nixosTest nodes get the qemu-vm module automatically), so this also adds an explicit import of it. Documented the underlying resource cost (chromium is a full browser per PDF request, pm.max_children=8 means up to 8 can run concurrently) in CLAUDE.md as an operational note for real deployments, not just this test VM. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
47 lines
2.2 KiB
Nix
47 lines
2.2 KiB
Nix
{ proxyServer ? "nginx" }:
|
|
{ pkgs, modulesPath, ... }:
|
|
|
|
{
|
|
imports = [
|
|
(modulesPath + "/profiles/qemu-guest.nix")
|
|
# Not imported by default for a plain nixosSystem in this nixpkgs (only nixosTest nodes get it
|
|
# automatically) - needed explicitly so virtualisation.cores/memorySize below are real options.
|
|
(modulesPath + "/virtualisation/qemu-vm.nix")
|
|
(import ./common.nix { inherit proxyServer; })
|
|
];
|
|
|
|
# Required to build pkgs.chromium, which the module pulls in for PDF generation.
|
|
nixpkgs.config.allowUnfree = true;
|
|
|
|
# `nix flake check` evaluates config.system.build.toplevel directly, which asserts a real root
|
|
# filesystem and bootloader are configured. `just build-vm` builds config.system.build.vm
|
|
# instead, where the qemu-vm module overrides both of these with a tmpfs root and direct
|
|
# kernel boot (virtualisation.useBootLoader defaults to false) - so these values just need to
|
|
# satisfy the assertions and are otherwise unused.
|
|
fileSystems."/" = {
|
|
device = "/dev/disk/by-label/nixos";
|
|
fsType = "ext4";
|
|
};
|
|
boot.loader.grub.enable = false;
|
|
|
|
users.users.root = {
|
|
initialPassword = "toor";
|
|
};
|
|
|
|
services.resolved.enable = true;
|
|
|
|
# The qemu-vm module's defaults (1 core, 1024MB) are too tight for real interactive use: MySQL +
|
|
# Redis + PHP-FPM + chromium (PDF generation, spawned per request - each is a full multi-process
|
|
# browser, not a lightweight renderer) running concurrently can exhaust 1GB fast, and this VM
|
|
# boots with panic_on_oom - so it's not a graceful slowdown, the whole VM goes down. Confirmed by
|
|
# reproducing a "logo upload hangs" report this way: at the 1024MB default, even a couple of
|
|
# concurrent PDF-generation requests triggered `redis-server invoked oom-killer` and a kernel
|
|
# panic. 2 cores / 2560MB comfortably handles 2 concurrent PDF generations alongside normal
|
|
# traffic (verified: ~1.4GB still free afterward); a deliberate flood of many concurrent PDF
|
|
# requests can still exhaust even this, since each is a full chromium instance - if that's a
|
|
# concern for your deployment, consider capping `phpfpm.settings."pm.max_children"` below its
|
|
# default of 8, or provisioning more RAM.
|
|
virtualisation.cores = 2;
|
|
virtualisation.memorySize = 2560;
|
|
}
|