Files
sanguo_vnpy_v2/scripts/nas_sync/sync_parquet_resumable.sh

53 lines
2.1 KiB
Bash

#!/bin/bash
# NAS 端:VPS parquet(static+valuation) 文件级断点续传。
# VPS 无 rsync,改逐文件 scp + NAS 端 size 校验跳过已传。中断重跑只续未传文件。
# 脱离 ssh 会话: nohup setsid bash sync_parquet_resumable.sh </dev/null >/dev/null 2>&1 &
set -u
KEY=/var/services/homes/admin/.ssh/id_ed25519_nas
VPS=Administrator@49.232.102.198
NAS_DATA=/volume1/stock/sanguo_vnpy_v2/data
LOG=/volume1/stock/sanguo_vnpy_v2/data_backup/parquet.log
MANIFEST=/volume1/stock/sanguo_vnpy_v2/data_backup/parquet_manifest.txt
echo "=== resumable start $(date) ===" >> "$LOG"
# 1. VPS 端 Python 列 parquet 清单(path size, 正斜杠), stdin 喂执行不落盘
ssh -i "$KEY" -o StrictHostKeyChecking=no "$VPS" 'C:\Python310\python.exe -X utf8 -' > "$MANIFEST" <<'PYEOF'
import os
for sub in ("static", "valuation_baostock"):
root = os.path.join(r"C:\sanguo_vnpy_v2\data", sub)
for dp, _, fs in os.walk(root):
for f in fs:
p = os.path.join(dp, f).replace("\\", "/")
try:
print(p, os.path.getsize(p))
except OSError:
pass
PYEOF
n=$(wc -l < "$MANIFEST")
echo "[manifest] $n files" >> "$LOG"
# 2. 逐文件 scp, NAS size 匹配则跳过(断点续传核心)
done=0; skipped=0; failed=0
while read -r vps_path size rest; do
[ -z "$vps_path" ] && continue
rel=${vps_path#C:/sanguo_vnpy_v2/data/}
nas_path="$NAS_DATA/$rel"
mkdir -p "$(dirname "$nas_path")"
if [ -f "$nas_path" ] && [ "$(stat -c%s "$nas_path" 2>/dev/null)" = "$size" ]; then
skipped=$((skipped+1))
elif scp -q -i "$KEY" -o StrictHostKeyChecking=no "$VPS:$vps_path" "$nas_path" 2>>"$LOG"; then
done=$((done+1))
if [ $((done % 50)) -eq 0 ]; then echo "[progress] done=$done skipped=$skipped failed=$failed" >> "$LOG"; fi
else
failed=$((failed+1))
echo "[FAIL] $rel" >> "$LOG"
fi
done < "$MANIFEST"
echo "[result] done=$done skipped=$skipped failed=$failed" >> "$LOG"
du -sh "$NAS_DATA/static" "$NAS_DATA/valuation_baostock" >> "$LOG" 2>&1
echo "=== resumable done $(date) ===" >> "$LOG"
echo "PARQUET_DONE $(date)" >> "$LOG"