增加 payload replay for snapshot，20260313 迭代

2026-03-15 22:49:06 +08:00
parent cf764c35bb
commit 73d8ebb5c1
17 changed files with 3060 additions and 21 deletions
--- a/scripts/payload_replay/README.md
+++ b/scripts/payload_replay/README.md
@@ -0,0 +1,100 @@
+# Payload Replay Scripts
+
+本目录提供基于本地 payload archive 的手工 replay 入口。
+
+## `run_apnic_replay.sh`
+
+默认使用：
+
+- `tests/fixtures/tal/apnic-rfc7730-https.tal`
+- `tests/fixtures/ta/apnic-ta.cer`
+- `target/live/payload_replay/payload-archive`
+- `target/live/payload_replay/locks.json`
+
+运行：
+
+```bash
+./scripts/payload_replay/run_apnic_replay.sh
+```
+
+产物默认输出到：
+
+- `target/live/payload_replay_runs/`
+
+包含：
+
+- replay DB 目录
+- `report.json`
+- `run.log`
+- `meta.json`
+- `summary.md`
+
+## 环境变量
+
+可覆盖：
+
+- `TAL_PATH`
+- `TA_PATH`
+- `PAYLOAD_REPLAY_ARCHIVE`
+- `PAYLOAD_REPLAY_LOCKS`
+- `VALIDATION_TIME`
+- `MAX_DEPTH`
+- `MAX_INSTANCES`
+- `OUT_DIR`
+- `RUN_NAME`
+- `DB_DIR`
+- `REPORT_JSON`
+- `RUN_LOG`
+- `META_JSON`
+- `SUMMARY_MD`
+
+## 说明
+
+- 该脚本依赖 `rpki` CLI 已支持：
+  - `--payload-replay-archive`
+  - `--payload-replay-locks`
+- replay 模式必须搭配离线 TAL/TA 输入，不会去访问真实 RRDP / rsync 网络源。
+
+## `report_to_routinator_csv.py`
+
+把 `rpki` 生成的 `report.json` 转成 Routinator 风格的 VRP CSV：
+
+```bash
+python3 scripts/payload_replay/report_to_routinator_csv.py \
+  --report target/live/payload_replay_runs/<run>_report.json \
+  --out target/live/payload_replay_runs/<run>_vrps.csv \
+  --trust-anchor apnic
+```
+
+输出列为：
+
+- `ASN`
+- `IP Prefix`
+- `Max Length`
+- `Trust Anchor`
+
+## `compare_with_routinator_record.sh`
+
+把 ours 生成的 VRP CSV 与 Routinator 的 `record.csv` 做对比：
+
+```bash
+./scripts/payload_replay/compare_with_routinator_record.sh \
+  target/live/payload_replay_runs/<run>_vrps.csv \
+  target/live/payload_replay/record.csv
+```
+
+会产出：
+
+- compare summary Markdown
+- `only_in_ours.csv`
+- `only_in_record.csv`
+
+## `run_apnic_replay.sh` 现有额外产物
+
+脚本现在除了 `report/meta/summary`，还会额外生成：
+
+- `vrps.csv`
+- 若 `ROUTINATOR_RECORD_CSV` 存在，则生成：
+  - compare summary
+  - `only_in_ours.csv`
+  - `only_in_record.csv`
--- a/scripts/payload_replay/compare_with_routinator_record.sh
+++ b/scripts/payload_replay/compare_with_routinator_record.sh
@@ -0,0 +1,110 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+if [[ $# -lt 2 || $# -gt 5 ]]; then
+  echo "Usage: $0 <ours.csv> <record.csv> [summary.md] [only_in_ours.csv] [only_in_record.csv]" >&2
+  exit 2
+fi
+
+OURS_CSV="$1"
+RECORD_CSV="$2"
+SUMMARY_MD="${3:-}"
+ONLY_IN_OURS_CSV="${4:-}"
+ONLY_IN_RECORD_CSV="${5:-}"
+
+if [[ -z "$SUMMARY_MD" ]]; then
+  SUMMARY_MD="$(dirname "$OURS_CSV")/$(basename "$OURS_CSV" .csv)_vs_routinator_summary.md"
+fi
+if [[ -z "$ONLY_IN_OURS_CSV" ]]; then
+  ONLY_IN_OURS_CSV="$(dirname "$OURS_CSV")/$(basename "$OURS_CSV" .csv)_only_in_ours.csv"
+fi
+if [[ -z "$ONLY_IN_RECORD_CSV" ]]; then
+  ONLY_IN_RECORD_CSV="$(dirname "$OURS_CSV")/$(basename "$OURS_CSV" .csv)_only_in_record.csv"
+fi
+
+python3 - "$OURS_CSV" "$RECORD_CSV" "$SUMMARY_MD" "$ONLY_IN_OURS_CSV" "$ONLY_IN_RECORD_CSV" <<'PY'
+import csv
+import ipaddress
+import sys
+from pathlib import Path
+
+ours_csv = Path(sys.argv[1])
+record_csv = Path(sys.argv[2])
+summary_md = Path(sys.argv[3])
+only_in_ours_csv = Path(sys.argv[4])
+only_in_record_csv = Path(sys.argv[5])
+
+
+def normalize_row(row: dict):
+    asn = row["ASN"].strip().upper()
+    prefix = row["IP Prefix"].strip()
+    max_len = str(int(row["Max Length"]))
+    ta = row["Trust Anchor"].strip()
+    network = ipaddress.ip_network(prefix, strict=False)
+    return {
+        "ASN": asn,
+        "IP Prefix": str(network),
+        "Max Length": max_len,
+        "Trust Anchor": ta,
+    }
+
+
+def read_rows(path: Path):
+    with path.open(encoding="utf-8", newline="") as f:
+        rows = [normalize_row(r) for r in csv.DictReader(f)]
+    return rows
+
+
+def row_key(row: dict):
+    network = ipaddress.ip_network(row["IP Prefix"], strict=False)
+    return (
+        row["ASN"],
+        network.version,
+        int(network.network_address),
+        network.prefixlen,
+        int(row["Max Length"]),
+        row["Trust Anchor"],
+    )
+
+
+def write_rows(path: Path, rows):
+    path.parent.mkdir(parents=True, exist_ok=True)
+    with path.open("w", encoding="utf-8", newline="") as f:
+        writer = csv.DictWriter(f, fieldnames=["ASN", "IP Prefix", "Max Length", "Trust Anchor"])
+        writer.writeheader()
+        for row in rows:
+            writer.writerow(row)
+
+ours = read_rows(ours_csv)
+record = read_rows(record_csv)
+ours_map = {row_key(r): r for r in ours}
+record_map = {row_key(r): r for r in record}
+only_in_ours = [ours_map[k] for k in sorted(set(ours_map) - set(record_map))]
+only_in_record = [record_map[k] for k in sorted(set(record_map) - set(ours_map))]
+intersection = len(set(ours_map) & set(record_map))
+
+write_rows(only_in_ours_csv, only_in_ours)
+write_rows(only_in_record_csv, only_in_record)
+
+summary_md.parent.mkdir(parents=True, exist_ok=True)
+summary = []
+summary.append("# Replay vs Routinator VRP Compare\n\n")
+summary.append(f"- ours_csv: `{ours_csv}`\n")
+summary.append(f"- record_csv: `{record_csv}`\n")
+summary.append(f"- only_in_ours_csv: `{only_in_ours_csv}`\n")
+summary.append(f"- only_in_record_csv: `{only_in_record_csv}`\n\n")
+summary.append("| metric | value |\n")
+summary.append("|---|---:|\n")
+summary.append(f"| ours_total | {len(ours_map)} |\n")
+summary.append(f"| record_total | {len(record_map)} |\n")
+summary.append(f"| intersection | {intersection} |\n")
+summary.append(f"| only_in_ours | {len(only_in_ours)} |\n")
+summary.append(f"| only_in_record | {len(only_in_record)} |\n")
+summary_md.write_text("".join(summary), encoding="utf-8")
+print(summary_md)
+PY
+
+echo "== compare complete ==" >&2
+echo "- summary: $SUMMARY_MD" >&2
+echo "- only_in_ours: $ONLY_IN_OURS_CSV" >&2
+echo "- only_in_record: $ONLY_IN_RECORD_CSV" >&2
--- a/scripts/payload_replay/report_to_routinator_csv.py
+++ b/scripts/payload_replay/report_to_routinator_csv.py
@@ -0,0 +1,57 @@
+#!/usr/bin/env python3
+import argparse
+import csv
+import ipaddress
+import json
+from pathlib import Path
+
+
+def parse_args() -> argparse.Namespace:
+    p = argparse.ArgumentParser(
+        description="Convert rpki report.json VRPs into Routinator-compatible CSV"
+    )
+    p.add_argument("--report", required=True, help="path to rpki report.json")
+    p.add_argument("--out", required=True, help="output CSV path")
+    p.add_argument(
+        "--trust-anchor",
+        default="unknown",
+        help="Trust Anchor column value (default: unknown)",
+    )
+    return p.parse_args()
+
+
+def sort_key(vrp: dict):
+    network = ipaddress.ip_network(vrp["prefix"], strict=False)
+    return (
+        int(vrp["asn"]),
+        network.version,
+        int(network.network_address),
+        network.prefixlen,
+        int(vrp["max_length"]),
+    )
+
+
+def main() -> int:
+    args = parse_args()
+    report = json.loads(Path(args.report).read_text(encoding="utf-8"))
+    vrps = list(report.get("vrps") or [])
+    vrps.sort(key=sort_key)
+
+    out_path = Path(args.out)
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+    with out_path.open("w", encoding="utf-8", newline="") as f:
+        w = csv.writer(f)
+        w.writerow(["ASN", "IP Prefix", "Max Length", "Trust Anchor"])
+        for vrp in vrps:
+            w.writerow([
+                f"AS{vrp['asn']}",
+                vrp["prefix"],
+                vrp["max_length"],
+                args.trust_anchor,
+            ])
+    print(out_path)
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
--- a/scripts/payload_replay/run_apnic_replay.sh
+++ b/scripts/payload_replay/run_apnic_replay.sh
@@ -0,0 +1,150 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
+cd "$ROOT_DIR"
+
+TAL_PATH="${TAL_PATH:-$ROOT_DIR/tests/fixtures/tal/apnic-rfc7730-https.tal}"
+TA_PATH="${TA_PATH:-$ROOT_DIR/tests/fixtures/ta/apnic-ta.cer}"
+PAYLOAD_REPLAY_ARCHIVE="${PAYLOAD_REPLAY_ARCHIVE:-$ROOT_DIR/target/live/payload_replay/payload-archive}"
+PAYLOAD_REPLAY_LOCKS="${PAYLOAD_REPLAY_LOCKS:-$ROOT_DIR/target/live/payload_replay/locks.json}"
+VALIDATION_TIME="${VALIDATION_TIME:-2026-03-13T02:30:00Z}"
+TRUST_ANCHOR="${TRUST_ANCHOR:-apnic}"
+ROUTINATOR_RECORD_CSV="${ROUTINATOR_RECORD_CSV:-$ROOT_DIR/target/live/payload_replay/record.csv}"
+MAX_DEPTH="${MAX_DEPTH:-}"
+MAX_INSTANCES="${MAX_INSTANCES:-}"
+OUT_DIR="${OUT_DIR:-$ROOT_DIR/target/live/payload_replay_runs}"
+mkdir -p "$OUT_DIR"
+
+TS="$(date -u +%Y%m%dT%H%M%SZ)"
+RUN_NAME="${RUN_NAME:-apnic_replay_${TS}}"
+DB_DIR="${DB_DIR:-$OUT_DIR/${RUN_NAME}_db}"
+REPORT_JSON="${REPORT_JSON:-$OUT_DIR/${RUN_NAME}_report.json}"
+RUN_LOG="${RUN_LOG:-$OUT_DIR/${RUN_NAME}_run.log}"
+META_JSON="${META_JSON:-$OUT_DIR/${RUN_NAME}_meta.json}"
+SUMMARY_MD="${SUMMARY_MD:-$OUT_DIR/${RUN_NAME}_summary.md}"
+VRPS_CSV="${VRPS_CSV:-$OUT_DIR/${RUN_NAME}_vrps.csv}"
+COMPARE_SUMMARY_MD="${COMPARE_SUMMARY_MD:-$OUT_DIR/${RUN_NAME}_compare_summary.md}"
+ONLY_IN_OURS_CSV="${ONLY_IN_OURS_CSV:-$OUT_DIR/${RUN_NAME}_only_in_ours.csv}"
+ONLY_IN_RECORD_CSV="${ONLY_IN_RECORD_CSV:-$OUT_DIR/${RUN_NAME}_only_in_record.csv}"
+
+cmd=(cargo run --release --bin rpki --
+  --db "$DB_DIR"
+  --tal-path "$TAL_PATH"
+  --ta-path "$TA_PATH"
+  --payload-replay-archive "$PAYLOAD_REPLAY_ARCHIVE"
+  --payload-replay-locks "$PAYLOAD_REPLAY_LOCKS"
+  --validation-time "$VALIDATION_TIME"
+  --report-json "$REPORT_JSON")
+
+if [[ -n "$MAX_DEPTH" ]]; then
+  cmd+=(--max-depth "$MAX_DEPTH")
+fi
+if [[ -n "$MAX_INSTANCES" ]]; then
+  cmd+=(--max-instances "$MAX_INSTANCES")
+fi
+
+run_start_s="$(date +%s)"
+(
+  echo "# command:"
+  printf '%q ' "${cmd[@]}"
+  echo
+  echo
+  "${cmd[@]}"
+) 2>&1 | tee "$RUN_LOG" >/dev/null
+run_end_s="$(date +%s)"
+run_duration_s="$((run_end_s - run_start_s))"
+
+TAL_PATH="$TAL_PATH" \
+TA_PATH="$TA_PATH" \
+PAYLOAD_REPLAY_ARCHIVE="$PAYLOAD_REPLAY_ARCHIVE" \
+PAYLOAD_REPLAY_LOCKS="$PAYLOAD_REPLAY_LOCKS" \
+DB_DIR="$DB_DIR" \
+REPORT_JSON="$REPORT_JSON" \
+RUN_LOG="$RUN_LOG" \
+VALIDATION_TIME="$VALIDATION_TIME" \
+RUN_DURATION_S="$run_duration_s" \
+python3 - "$REPORT_JSON" "$META_JSON" "$SUMMARY_MD" <<'PY'
+import json
+import os
+import sys
+from datetime import datetime, timezone
+from pathlib import Path
+
+report_path = Path(sys.argv[1])
+meta_path = Path(sys.argv[2])
+summary_path = Path(sys.argv[3])
+rep = json.loads(report_path.read_text(encoding="utf-8"))
+now = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+meta = {
+  "recorded_at_utc": now,
+  "tal_path": os.environ["TAL_PATH"],
+  "ta_path": os.environ["TA_PATH"],
+  "payload_replay_archive": os.environ["PAYLOAD_REPLAY_ARCHIVE"],
+  "payload_replay_locks": os.environ["PAYLOAD_REPLAY_LOCKS"],
+  "db_dir": os.environ["DB_DIR"],
+  "report_json": os.environ["REPORT_JSON"],
+  "run_log": os.environ["RUN_LOG"],
+  "validation_time_arg": os.environ["VALIDATION_TIME"],
+  "durations_secs": {
+    "rpki_run": int(os.environ["RUN_DURATION_S"]),
+  },
+  "counts": {
+    "publication_points_processed": rep["tree"]["instances_processed"],
+    "publication_points_failed": rep["tree"]["instances_failed"],
+    "vrps": len(rep["vrps"]),
+    "aspas": len(rep["aspas"]),
+    "audit_publication_points": len(rep["publication_points"]),
+  },
+}
+meta_path.write_text(json.dumps(meta, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
+summary = []
+summary.append("# Payload Replay Summary\n\n")
+summary.append(f"- recorded_at_utc: `{now}`\n")
+summary.append(f"- tal_path: `{meta['tal_path']}`\n")
+summary.append(f"- ta_path: `{meta['ta_path']}`\n")
+summary.append(f"- payload_replay_archive: `{meta['payload_replay_archive']}`\n")
+summary.append(f"- payload_replay_locks: `{meta['payload_replay_locks']}`\n")
+summary.append(f"- db: `{meta['db_dir']}`\n")
+summary.append(f"- report_json: `{meta['report_json']}`\n")
+summary.append(f"- validation_time_arg: `{meta['validation_time_arg']}`\n\n")
+summary.append("## Results\n\n")
+summary.append("| metric | value |\n")
+summary.append("|---|---:|\n")
+for k, v in meta["counts"].items():
+    summary.append(f"| {k} | {v} |\n")
+summary.append("\n## Durations\n\n")
+summary.append("| step | seconds |\n")
+summary.append("|---|---:|\n")
+for k, v in meta["durations_secs"].items():
+    summary.append(f"| {k} | {v} |\n")
+summary_path.write_text("".join(summary), encoding="utf-8")
+print(summary_path)
+PY
+
+python3 scripts/payload_replay/report_to_routinator_csv.py \
+  --report "$REPORT_JSON" \
+  --out "$VRPS_CSV" \
+  --trust-anchor "$TRUST_ANCHOR" >/dev/null
+
+if [[ -f "$ROUTINATOR_RECORD_CSV" ]]; then
+  ./scripts/payload_replay/compare_with_routinator_record.sh \
+    "$VRPS_CSV" \
+    "$ROUTINATOR_RECORD_CSV" \
+    "$COMPARE_SUMMARY_MD" \
+    "$ONLY_IN_OURS_CSV" \
+    "$ONLY_IN_RECORD_CSV" >/dev/null
+fi
+
+echo "== payload replay run complete ==" >&2
+echo "- db: $DB_DIR" >&2
+echo "- report: $REPORT_JSON" >&2
+echo "- run log: $RUN_LOG" >&2
+echo "- meta json: $META_JSON" >&2
+echo "- summary md: $SUMMARY_MD" >&2
+echo "- vrps csv: $VRPS_CSV" >&2
+if [[ -f "$COMPARE_SUMMARY_MD" ]]; then
+  echo "- compare summary: $COMPARE_SUMMARY_MD" >&2
+  echo "- only in ours: $ONLY_IN_OURS_CSV" >&2
+  echo "- only in record: $ONLY_IN_RECORD_CSV" >&2
+fi