# 启动 scrub（后台运行）
sudo btrfs scrub start /mnt/data

# 启动并等待完成
sudo btrfs scrub start -B /mnt/data

# 只 scrub 指定设备
sudo btrfs scrub start -d /dev/sdb /mnt/data

# 限制 scrub 速度（降低对业务的影响）
sudo btrfs scrub start -c 2 /mnt/data  # 限制 2 个并发

# 设置 IO 优先级（需要 ionice 支持）
sudo ionice -c 3 btrfs scrub start /mnt/data

# 查看当前 scrub 状态
sudo btrfs scrub status /mnt/data

# 输出示例：
# UUID:             a1b2c3d4-e5f6-7890-abcd-ef1234567890
# Scrub started:    Sat May 10 02:00:00 2026
# Status:           running
# Duration:         0:35:22
# Total to scrub:   100.00GiB
# Rate:             48.50MiB/s
# Error summary:    no errors found

# 查看历史 scrub 结果
sudo btrfs scrub status -d /mnt/data

# 取消正在进行的 scrub
sudo btrfs scrub cancel /mnt/data

# 查看 scrub 错误详情
sudo btrfs scrub status -d /mnt/data

# 输出中错误示例：
# Scrub device /dev/sdb (id 1) history
# Scrub started:    Sat May 10 02:00:00 2026
# Status:           finished
# Duration:         0:35:22
# Total to scrub:   100.00GiB
# Rate:             48.50MiB/s
# Error summary:    csum=3
#   Corrected:      2
#   Uncorrectable:  1
#   Unverified:     0

# 查看设备错误统计
sudo btrfs device stats /mnt/data

# 如果 scrub 发现错误且无法修复
# 1. 立即备份数据
sudo rsync -av /mnt/data/ /backup/

# 2. 检查文件系统
sudo umount /mnt/data
sudo btrfs check /dev/sdb1

# 3. 如果 check 报错
sudo btrfs check --repair /dev/sdb1  # ⚠️ 危险操作

# RAID 1 模式下，scrub 会自动从其他副本修复损坏的数据
# scrub 结果中 "Corrected" 数量会增加

# 查看修复详情
sudo btrfs device stats /mnt/data
# [/dev/sdb].corruption_errs  3   ← 检测到 3 个损坏
# [/dev/sdc].corruption_errs  0

# 修复后的数据已经写回正确的副本

# /etc/systemd/system/btrfs-scrub@.timer
[Unit]
Description=Btrfs scrub timer for %i

[Timer]
OnCalendar=weekly
RandomizedDelaySec=3600
Persistent=true

[Install]
WantedBy=timers.target

# /etc/systemd/system/btrfs-scrub@.service
[Unit]
Description=Btrfs scrub for %i

[Service]
Type=oneshot
ExecStart=/usr/bin/btrfs scrub start -B %i
Nice=19
IOSchedulingClass=idle

# 启用 scrub 定时器
sudo systemctl enable --now btrfs-scrub@mnt-data.timer

# 检查定时器状态
sudo systemctl list-timers btrfs-scrub@mnt-data.timer

# 每周日凌晨 2 点执行 scrub
0 2 * * 0 /usr/bin/btrfs scrub start -B /data 2>&1 | logger -t btrfs-scrub

# 每月 1 号凌晨 3 点（SSD）
0 3 1 * * /usr/bin/btrfs scrub start -B /data 2>&1 | logger -t btrfs-scrub

#!/bin/bash
# btrfs-scrub-wrapper.sh - 带通知的 Scrub 脚本
set -euo pipefail

MOUNT_POINT="${1:?Usage: $0 /mount/point}"
ALERT_EMAIL="${2:-admin@example.com}"
LOG_FILE="/var/log/btrfs-scrub.log"

log() {
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $*" | tee -a "$LOG_FILE"
}

send_alert() {
    local subject="$1"
    local body="$2"
    echo "$body" | mail -s "Btrfs Scrub: $subject" "$ALERT_EMAIL"
}

log "Starting scrub on $MOUNT_POINT"

# 执行 scrub
OUTPUT=$(sudo btrfs scrub start -B "$MOUNT_POINT" 2>&1) || true
log "$OUTPUT"

# 检查结果
STATUS=$(sudo btrfs scrub status "$MOUNT_POINT" 2>&1)
log "$STATUS"

# 检查错误统计
ERRORS=$(sudo btrfs device stats "$MOUNT_POINT" 2>&1)
TOTAL_ERRORS=$(echo "$ERRORS" | grep -v " 0$" | grep -v "^[/" | wc -l)

if [[ "$TOTAL_ERRORS" -gt 0 ]]; then
    log "WARNING: Device errors detected!"
    log "$ERRORS"
    send_alert "Scrub completed with errors on $MOUNT_POINT" \
        "Scrub output:\n$OUTPUT\n\nDevice stats:\n$ERRORS"
else
    log "Scrub completed successfully, no errors found."
fi

# 方法 1：降低优先级
sudo ionice -c 3 btrfs scrub start /mnt/data
sudo nice -n 19 btrfs scrub start /mnt/data

# 方法 2：限制并发设备数
sudo btrfs scrub start -c 1 /mnt/data

# 方法 3：选择低峰期执行
# cron 中设置在凌晨执行
0 2 * * 0 /usr/bin/btrfs scrub start -B /data

# scrub 会校验所有数据，包括快照中的数据
# 快照与源子卷共享数据块，所以 scrub 高效地一次验证

# RAID 1：scrub 自动修复损坏的副本
sudo btrfs scrub start /mnt/raid1
# 错误会被自动修复

# RAID 5/6：scrub 可以使用奇偶校验修复
# 但由于 write hole 问题，修复可能不完全可靠