packaging/windows: fix BTRFS+QEMU boot-loop, document what's confirmed #3

Merged
dodox merged 1 commits from fix/id-space-strike-request-perf-2026-08-11 into master 2026-08-11 21:17:18 +02:00
2 changed files with 33 additions and 6 deletions
Showing only changes of commit 2282a7d521 - Show all commits

View File

@ -6,12 +6,28 @@ machine and no GitHub, by booting a real Windows VM inside a container
hood, no license key needed for the eval install it fetches automatically)
and driving the whole build over a shared folder.
**Status: written, not yet run against a real boot.** Everything here
follows dockur/windows's and WiX's documented mechanics, but there's no
KVM/Windows available in the environment this was authored in to actually
exercise it end to end. Treat the first run as a debugging session, not a
push-button success — watch it happen at http://localhost:8006 (dockur's
noVNC viewer) so you can see where it's stuck if it stalls.
**Status: has actually been run.** Two real environment issues hit and
fixed so far (both already applied in this directory, see "Confirmed
issues hit" below); Windows install itself was still in progress as of
last check. `oem/install.bat` onward (MSYS2/GTK4/WiX provisioning, the
actual .msi packaging) has NOT been reached/exercised yet — treat that
part as still a debugging session, not a push-button success. Watch it
happen at http://localhost:8006 (dockur's noVNC viewer) if it stalls.
## Confirmed issues hit (on Fedora + BTRFS)
- **SELinux blocks the bind mounts** ("Storage folder (/storage) is not
writeable!") — fixed with `:Z` on every volume in `docker-compose.yml`.
Harmless no-op on a host without SELinux.
- **BTRFS + QEMU disk images boot-loop Windows Setup for hours** with no
error, just the same boot-manager log lines repeating forever and the
disk barely growing — dockur/windows warns about this itself
("you are using the BTRFS filesystem for /storage, this might
introduce issues with Windows Setup!"), and on this host it wasn't
idle chatter. Fixed by disabling copy-on-write on `storage/` before
the VM ever writes to it (`chattr +C`, only affects files created
after it's set, so it needs an empty directory) — `build_windows.sh`
now does this itself on every run. No-op on a non-btrfs filesystem.
## How it fits together

View File

@ -30,6 +30,17 @@ fi
command -v docker >/dev/null 2>&1 || { echo "docker not found." >&2; exit 1; }
mkdir -p storage oem shared/src shared/dist "$OUT_DIR"
# dockur/windows itself warns about this ("you are using the BTRFS
# filesystem for /storage, this might introduce issues with Windows
# Setup!") and it's not idle: confirmed on this host as a genuine
# multi-hour Windows Setup boot-loop (repeating the same boot-manager
# log lines forever, disk barely growing) -- QEMU disk images on a
# copy-on-write filesystem are a known bad combination. +C only takes
# effect for files created AFTER it's set on an empty directory, so
# this only helps on a fresh/emptied storage/; it's a no-op (harmless,
# chattr just errors quietly) on a non-btrfs filesystem or an
# already-populated storage/ from a previous run.
chattr +C storage 2>/dev/null || true
echo "==> starting the Windows build VM (docker compose up -d)"
docker compose up -d