ykxmz/_build/rp_ab_long.sh

#!/bin/bash
# 长窗口配比对照用的测量脚本（支持 A/B 两组，各跑 >=40 分钟）
#
# 与 rp_long2h.sh 的区别：
#   * 时长可指定（分钟）
#   * **额外采样内存和 CPU**（Lead 问 h2 提到 4 条对 460MB 设备有无影响）
#   * 输出文件名带标签，方便 A/B 分开比对
#
# 用法: bash rp_ab_long.sh <标签> <分钟数>
set -u

LABEL=${1:-A}
MINS=${2:-45}
BASE=https://relay.xxmxxs.top
UA='LanRemote/1.8.2'
DUR=$((MINS * 60))
HEALTH_EVERY=15
CONN_EVERY=60
RES_EVERY=60

H=/root/ab_${LABEL}_health.tsv
C=/root/ab_${LABEL}_conns.tsv
R=/root/ab_${LABEL}_res.tsv
S=/root/ab_${LABEL}_summary.txt

rm -f "$H" "$C" "$R" "$S"

conns_of() {
    curl -s --max-time 4 "http://127.0.0.1:$1/metrics" 2>/dev/null \
        | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}'
}

M0=$(wc -l < /root/lanremote-relay/tunnel.log 2>/dev/null || echo 0)
H0=$(wc -l < /root/lanremote-relay/tunnel-h2.log 2>/dev/null || echo 0)

T0=$(date +%s)
NEXT_CONN=$T0
NEXT_RES=$T0

echo "  开始 $LABEL: $(date '+%H:%M:%S')  持续 ${MINS} 分钟"

while true; do
    NOW=$(date +%s)
    [ $((NOW - T0)) -ge "$DUR" ] && break

    out=$(curl -s -o /dev/null -w '%{http_code} %{time_total}' --max-time 25 \
          -H "User-Agent: $UA" "$BASE/v1/health")
    code=$(echo "$out" | awk '{print $1}')
    ms=$(echo "$out" | awk '{printf "%d", $2*1000}')
    echo "$(date '+%Y-%m-%d %H:%M:%S')	$code	$ms" >> "$H"

    if [ "$NOW" -ge "$NEXT_CONN" ]; then
        a=$(conns_of 20241); b=$(conns_of 20242)
        echo "$(date '+%Y-%m-%d %H:%M:%S')	${a:-X}	${b:-X}" >> "$C"
        NEXT_CONN=$((NOW + CONN_EVERY))
    fi

    if [ "$NOW" -ge "$NEXT_RES" ]; then
        # 两个 cloudflared 各占多少常驻内存 + CPU
        qm=0; hm=0; qc=0; hc=0
        for p in $(pgrep cloudflared); do
            proto=$(tr '\0' ' ' < /proc/$p/cmdline 2>/dev/null | grep -oE '\-\-protocol [a-z0-9]+' | awk '{print $2}')
            rss=$(awk '/VmRSS/{print $2}' /proc/$p/status 2>/dev/null)
            cpu=$(ps -o %cpu= -p $p 2>/dev/null | tr -d ' ')
            if [ "$proto" = "quic" ]; then qm=${rss:-0}; qc=${cpu:-0};
            elif [ "$proto" = "http2" ]; then hm=${rss:-0}; hc=${cpu:-0}; fi
        done
        total=$(free -m | awk '/Mem:/{print $3}')
        avail=$(free -m | awk '/Mem:/{print $7}')
        load=$(awk '{print $1}' /proc/loadavg)
        echo "$(date '+%Y-%m-%d %H:%M:%S')	${qm}	${hm}	${qc}	${hc}	${total}	${avail}	${load}" >> "$R"
        NEXT_RES=$((NOW + RES_EVERY))
    fi

    sleep "$HEALTH_EVERY"
done

M1=$(wc -l < /root/lanremote-relay/tunnel.log 2>/dev/null || echo 0)
H1=$(wc -l < /root/lanremote-relay/tunnel-h2.log 2>/dev/null || echo 0)

{
    echo "════════════════════════════════════════════════════════"
    echo "  $LABEL 组 · ${MINS} 分钟 · $(date '+%Y-%m-%d %H:%M:%S')"
    echo "════════════════════════════════════════════════════════"
    echo
    echo "── 配置 ──"
    for svc in lanremote-tunnel lanremote-tunnel-h2; do
        echo "  $svc: $(systemctl cat $svc 2>/dev/null | grep -oE '\-\-protocol [a-z0-9]+|\-\-ha-connections [0-9]+' | tr '\n' ' ')"
    done
    echo
    echo "── 公网可用率 ──"
    n=$(grep -c . "$H" 2>/dev/null || echo 0)
    ok=$(awk -F'\t' '$2=="200"' "$H" 2>/dev/null | wc -l)
    echo "  $ok/$n = $(awk -v a="$ok" -v b="$n" 'BEGIN{if(b>0)printf "%.2f", a*100/b; else print 0}')%"
    echo "  状态码: $(awk -F'\t' '{print $2}' "$H" 2>/dev/null | sort | uniq -c | tr '\n' ' ')"
    awk -F'\t' '$2=="200"{print $3}' "$H" 2>/dev/null | sort -n | awk '
        {a[NR]=$1}
        END{if(NR==0){print "  无数据"; exit}
            printf "  耗时: 最小 %d 中位 %d P90 %d P99 %d 最大 %d ms\n",
                a[1], a[int(NR/2)+1], a[int(NR*0.9)+1], a[int(NR*0.99)+1], a[NR]}'
    echo "  失败时刻:"
    awk -F'\t' '$2!="200"{print "    " $1 " 码=" $2 " " $3 "ms"}' "$H" 2>/dev/null
    echo
    echo "── 连接数 ──"
    awk -F'\t' 'NR>1 && $2 ~ /^[0-9]+$/ && $3 ~ /^[0-9]+$/ {
        t=$2+$3; cnt[t]++
        if (t<=3) low++
        if (t<min || min=="") min=t
        if (t>max) max=t
    } END {
        printf "  最小 %d 最大 %d 平均 %.2f\n", min, max, 0
        printf "  总连接数<=3 的采样: %d 次（占 %.1f%%）\n", low+0, (low+0)*100/NR
    }' "$C" 2>/dev/null
    echo "  分布: $(awk -F'\t' 'NR>1 && $2 ~ /^[0-9]+$/ && $3 ~ /^[0-9]+$/ {print $2+$3}' "$C" | sort -n | uniq -c | tr '\n' ' ')"
    echo
    echo "── 内存 / CPU ──"
    awk -F'\t' 'NR>1 {
        if ($2+0>qm) qm=$2
        if ($3+0>hm) hm=$3
        if ($4+0>qc) qc=$4
        if ($5+0>hc) hc=$5
        if ($6+0>tm) tm=$6
        if (minav=="" || $7+0<minav) minav=$7
        if ($8+0>ld) ld=$8
    } END {
        printf "  QUIC 实例常驻内存峰值: %.1f MB\n", qm/1024
        printf "  HTTP/2 实例常驻内存峰值: %.1f MB\n", hm/1024
        printf "  两者合计: %.1f MB\n", (qm+hm)/1024
        printf "  CPU 峰值: quic %.1f%%  http2 %.1f%%\n", qc, hc
        printf "  设备已用内存峰值: %d MB   最少可用: %d MB\n", tm, minav
        printf "  负载峰值: %.2f\n", ld
    }' "$R" 2>/dev/null
    echo
    echo "── 隧道事件（本轮增量）──"
    printf "  QUIC   注册 %s  失败 %s  掉线 %s\n" \
        "$(tail -n +$((M0+1)) /root/lanremote-relay/tunnel.log 2>/dev/null | grep -c 'Registered tunnel connection')" \
        "$(tail -n +$((M0+1)) /root/lanremote-relay/tunnel.log 2>/dev/null | grep -c 'Failed to dial')" \
        "$(tail -n +$((M0+1)) /root/lanremote-relay/tunnel.log 2>/dev/null | grep -c 'Lost connection')"
    printf "  HTTP/2 注册 %s  失败 %s  掉线 %s\n" \
        "$(tail -n +$((H0+1)) /root/lanremote-relay/tunnel-h2.log 2>/dev/null | grep -c 'Registered tunnel connection')" \
        "$(tail -n +$((H0+1)) /root/lanremote-relay/tunnel-h2.log 2>/dev/null | grep -c 'Failed to dial')" \
        "$(tail -n +$((H0+1)) /root/lanremote-relay/tunnel-h2.log 2>/dev/null | grep -c 'Lost connection')"
    echo
    echo "── watchdog 处置（本轮）──"
    tail -12 /root/lanremote-relay/watchdog.log 2>/dev/null | sed 's/^/  /'
} > "$S" 2>&1

cat "$S"