blaze-host: check client sessions in BOTH network namespaces

Host-side ss cannot see the Python backend's connections: the responders run in
a container, so a client session terminates at 172.20.0.2:42130 inside its
namespace and the host only sees the NAT'd flow. 'ss | grep <client>' on the
host therefore reports nothing while a session is very much alive.

That produced a wrong precondition: 'no .105 Blaze session -- closed' was
reported while FIFA was mid-session on Python, and gate 9 was armed against a
client that had never exited. Python's own log had the answer -- it logs closes
reliably and there was no close for that session.

client-state.sh looks in both namespaces, reports Rust and Python separately,
and exits non-zero while any session is live. An unreachable container counts
as 'cannot confirm', not as 'clear'.

Fourth measurement bug in this tooling, and the most consequential: the other
three mis-COUNTED, this one mis-STATED a precondition and caused an action.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
funman300
2026-08-11 02:34:46 +00:00
parent b40adac3fc
commit ed0ccb8c2b
+81
View File
@@ -0,0 +1,81 @@
#!/usr/bin/env bash
# Is the FIFA client currently connected to anything?
#
# client-state.sh [client-ip]
#
# WHY THIS EXISTS
#
# Host-side `ss` CANNOT see the Python backend's connections. The Python
# responders run inside a container, so a client session terminates at
# 172.20.0.2:42130 in the container's network namespace; the host only ever sees
# the NAT'd flow, and a plain `ss | grep <client>` on the host reports nothing.
#
# That produced a wrong precondition check: "no .105 Blaze session — closed" was
# reported while FIFA was mid-session on Python, and gate 9 was armed against a
# client that had never exited. Python's own log gave it away — it logs closes
# reliably (45 of them) and there was no close for that session.
#
# So this looks in BOTH namespaces, and reports the Rust sidecar and the Python
# container separately.
#
# Exit 0 when the client has no live session anywhere (safe to start a gate),
# 1 when it does.
set -uo pipefail
CLIENT="${1:-${OPENFUT_CLIENT_IP:-10.10.0.105}}"
CONTAINER="${OPENFUT_PY_CONTAINER:-openfut-fut-backend}"
live=0
decode_tcp() {
# /proc/net/tcp rows -> "local remote state", little-endian hex addresses.
python3 -c "
import sys
def d(x):
ip, port = x.split(':')
return '.'.join(str(int(ip[i:i+2], 16)) for i in (6, 4, 2, 0)) + ':' + str(int(port, 16))
for line in sys.stdin:
f = line.split()
if len(f) < 4 or not f[0].endswith(':'):
continue
try:
print(d(f[1]), d(f[2]), f[3])
except Exception:
pass
"
}
echo "client: $CLIENT"
# ---- Rust sidecar (host namespace)
rust="$(ss -tn state established 2>/dev/null | grep -F "$CLIENT" | grep -E ':42230' || true)"
if [[ -n "$rust" ]]; then
echo " RUST sidecar : LIVE session(s)"
sed 's/^/ /' <<<"$rust"
live=1
else
echo " RUST sidecar : none"
fi
# ---- Python backend (container namespace)
if docker exec "$CONTAINER" true 2>/dev/null; then
py="$(docker exec "$CONTAINER" cat /proc/net/tcp 2>/dev/null | decode_tcp \
| awk '$3=="01"' | grep -F "$CLIENT" || true)"
if [[ -n "$py" ]]; then
echo " PYTHON backend: LIVE session(s)"
awk '{printf " %-22s <- %-22s ESTABLISHED\n", $1, $2}' <<<"$py"
live=1
else
echo " PYTHON backend: none"
fi
else
echo " PYTHON backend: container '$CONTAINER' not reachable — CANNOT confirm"
live=1 # unknown is not the same as clear
fi
echo
if [[ $live -eq 0 ]]; then
echo "RESULT: no live client session — safe to begin a gate"
else
echo "RESULT: client still connected — close FIFA before starting a gate"
fi
exit $live