fix: report only the retired services a host actually had #9

Merged
christianmanivong merged 1 commits from fix/retired-removal-truthful into main 2026-10-08 13:29:20 +00:00
3 changed files with 51 additions and 2 deletions

No files matched your search

+2 -1
View File
@@ -87,7 +87,8 @@ started earlier pulls whatever image the registry held before, which is stale.
4. Reconciles `apt-cacher-ng` and `signal-api`: each is recreated only if its
definition changed. It never touches `postgres` or `redis`.
It then removes services netOrk no longer ships, container and volume, where a host
still has them: today the bundled `registry` (NetOrk/netork#763).
still has them: today the bundled `registry` (NetOrk/netork#763). It reports only
what it actually removed; on a host that was cleaned earlier it says nothing.
5. Verifies that the containers run exactly the image that was pulled. If they don't, the
deploy fails.
6. Records `REGISTRY_HOST` and `NETORK_VERSION` in `~/netork/.env`, so that a
+5 -1
View File
@@ -316,6 +316,9 @@ deploy_steps() {
# file, so each would keep running on every host until someone removed it by
# hand. Idempotent: a host that never had one, or was cleaned already, passes
# untouched. Containers first; a volume still in use cannot be removed.
# Only what exists is removed and reported: `docker rm -f` exits 0 for a
# container that is not there, so without the check every deploy said it had
# removed the registry, long after it was gone (NetOrk/deploy#7).
# registry (netork-registry-1, netork_registry_data): the bundled registry:2
# for satellite images. Satellites pull from registry.netork.io; it held one
# old image, nothing had pulled from it in 30 days, and it accepted
@@ -325,7 +328,8 @@ deploy_steps() {
echo "[${SERVER}] Removing retired services..."
run_remote "$SERVER" "removing ${RETIRED_CONTAINERS} ${RETIRED_VOLUMES}" "" \
"for c in ${RETIRED_CONTAINERS}; do \
docker rm -f \$c >/dev/null 2>&1 && echo \"removed container \$c\"; done; \
docker inspect \$c >/dev/null 2>&1 && docker rm -f \$c >/dev/null 2>&1 \
&& echo \"removed container \$c\"; done; \
for v in ${RETIRED_VOLUMES}; do \
docker volume rm \$v >/dev/null 2>&1 && echo \"removed volume \$v\"; done; true" || true
+44
View File
@@ -71,3 +71,47 @@ def test_the_infrastructure_reconcile_no_longer_starts_the_registry(tmp_path: Pa
reconcile = [c for c in commands if "up -d --no-deps" in c and "--force-recreate" not in c]
assert reconcile, "the infrastructure reconcile step is missing"
assert " registry" not in reconcile[0].split("up -d --no-deps", 1)[1]
# ── It says it removed only what was there (NetOrk/deploy#7) ──────────────────
#
# `docker rm -f` exits 0 for a container that does not exist, so the step said
# "removed container netork-registry-1" on every deploy of every host, long after
# the container was gone -- which read as something recreating it each time.
_FAKE_DOCKER = """#!/bin/sh
case "$1" in
inspect) [ "{exists}" = yes ] ;;
rm) exit 0 ;;
volume) exit 1 ;;
esac
"""
def _removal_output(tmp_path: Path, *, exists: bool) -> str:
"""Run the deploy's removal command against a docker that has the container or not."""
_, commands = _deploy(tmp_path)
removal = next(c for c in commands if "netork-registry-1" in c)
docker_dir = tmp_path / "docker-bin"
docker_dir.mkdir()
docker = docker_dir / "docker"
docker.write_text(_FAKE_DOCKER.replace("{exists}", "yes" if exists else "no"))
docker.chmod(0o755)
run = subprocess.run(
["sh", "-c", removal],
env={"PATH": f"{docker_dir}:{os.environ['PATH']}"},
capture_output=True,
text=True,
timeout=30,
check=False,
)
assert run.returncode == 0, run.stderr
return run.stdout
def test_a_container_that_is_not_there_is_not_reported_removed(tmp_path: Path) -> None:
assert "removed container" not in _removal_output(tmp_path, exists=False)
def test_one_that_was_there_is(tmp_path: Path) -> None:
assert "removed container netork-registry-1" in _removal_output(tmp_path, exists=True)