# 检查当前内核版本
uname -r

# 对于生产环境，建议使用 LTS 内核
# Debian: apt install linux-image-amd64
# Ubuntu: 使用 HWE 内核
# Fedora: 使用最新稳定版
# Arch: 默认就是最新

#!/bin/bash
# production-format.sh

set -e

# 配置
DEVICE="/dev/sdb"
LABEL="data.0"
COMPRESSION="zstd"
COMPRESSION_LEVEL=3
CHECKSUM="crc32c"
BLOCK_SIZE="4k"

echo "=== 生产环境文件系统创建 ==="
echo ""
echo "设备: $DEVICE"
echo "压缩: $COMPRESSION (level $COMPRESSION_LEVEL)"
echo "校验和: $CHECKSUM"
echo "块大小: $BLOCK_SIZE"
echo ""

# 确认
read -p "这会清除 $DEVICE 上的所有数据！确认？(y/N) " confirm
if [ "$confirm" != "y" ]; then
    echo "取消"
    exit 0
fi

# 创建文件系统
sudo bcachefs format \
    --label="$LABEL" \
    --compression="$COMPRESSION" \
    --compression_level="$COMPRESSION_LEVEL" \
    --checksum="$CHECKSUM" \
    --block_size="$BLOCK_SIZE" \
    "$DEVICE"

# 验证
echo ""
echo "验证:"
sudo bcachefs show-super "$DEVICE"

echo ""
echo "=== 创建完成 ==="

# /etc/fstab 生产配置示例

# 单设备
UUID=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx /data bcachefs defaults,noatime,nodiratime,discard,compress=zstd 0 0

# 多设备
UUID=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx /data bcachefs defaults,noatime,nodiratime,discard,compress=zstd,device=/dev/sdb:/dev/sdc 0 0

# 挂载选项说明:
# noatime     - 不更新访问时间（性能优化）
# nodiratime  - 不更新目录访问时间
# discard     - SSD TRIM 支持
# compress    - 透明压缩

# 推荐的目录结构
/data/
├── app/              # 应用程序数据
├── database/         # 数据库文件
├── documents/        # 文档
├── logs/             # 日志
├── backups/          # 本地备份
├── snapshots/        # 快照目录
│   ├── daily/
│   ├── weekly/
│   └── monthly/
└── tmp/              # 临时文件

# 创建目录
sudo mkdir -p /data/{app,database,documents,logs,backups,snapshots/{daily,weekly,monthly},tmp}

#!/bin/bash
# bcachefs-monitor.sh

set -e

# 配置
MOUNT="/data"
LOG_FILE="/var/log/bcachefs-monitor.log"
ALERT_EMAIL="[email protected]"
SLACK_WEBHOOK=""  # 可选

# 函数: 发送告警
send_alert() {
    local level="$1"
    local message="$2"
    local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
    
    echo "[$timestamp] [$level] $message" >> "$LOG_FILE"
    
    # 发送邮件
    if [ -n "$ALERT_EMAIL" ]; then
        echo "$message" | mail -s "Bcachefs 告警 [$level]" "$ALERT_EMAIL" 2>/dev/null || true
    fi
    
    # 发送 Slack 通知
    if [ -n "$SLACK_WEBHOOK" ]; then
        curl -X POST -H 'Content-type: application/json' \
            --data "{\"text\":\"[$level] $message\"}" \
            "$SLACK_WEBHOOK" 2>/dev/null || true
    fi
}

# 检查挂载状态
if ! mountpoint -q "$MOUNT"; then
    send_alert "CRITICAL" "$MOUNT 未挂载"
    exit 1
fi

# 检查空间使用
USAGE=$(df --output=pcent "$MOUNT" | tail -1 | tr -d ' %')
if [ "$USAGE" -gt 95 ]; then
    send_alert "CRITICAL" "空间使用率 ${USAGE}%"
elif [ "$USAGE" -gt 85 ]; then
    send_alert "WARNING" "空间使用率 ${USAGE}%"
fi

# 检查内核错误
ERRORS=$(dmesg | grep -c "bcachefs.*error\|bcachefs.*corrupt" || echo 0)
if [ "$ERRORS" -gt 0 ]; then
    send_alert "WARNING" "发现 $ERRORS 个内核错误"
fi

# 检查设备状态
DEVICE_STATUS=$(sudo bcachefs device list "$MOUNT" 2>/dev/null || echo "")
if echo "$DEVICE_STATUS" | grep -q "failed\|offline"; then
    send_alert "CRITICAL" "发现故障设备"
fi

# 记录监控数据
echo "[$(date '+%Y-%m-%d %H:%M:%S')] usage=${USAGE}% errors=${ERRORS}" >> "$LOG_FILE"

# 自定义 Prometheus 指标收集器
#!/bin/bash
# /var/lib/node_exporter/textfile_collector/bcachefs.prom

MOUNT="/data"

# 空间使用率
USAGE=$(df --output=pcent "$MOUNT" | tail -1 | tr -d ' %')
echo "bcachefs_space_usage_percent $USAGE"

# 总空间
TOTAL=$(df -B1 --output=size "$MOUNT" | tail -1 | tr -d ' ')
echo "bcachefs_space_total_bytes $TOTAL"

# 已用空间
USED=$(df -B1 --output=used "$MOUNT" | tail -1 | tr -d ' ')
echo "bcachefs_space_used_bytes $USED"

# 错误计数
ERRORS=$(dmesg | grep -c "bcachefs.*error" || echo 0)
echo "bcachefs_kernel_errors_total $ERRORS"

# 使用 rsyslog 转发 bcachefs 日志
# /etc/rsyslog.d/bcachefs.conf

# 匹配 bcachefs 相关日志
:msg, contains, "bcachefs" /var/log/bcachefs.log
& stop

# 使用 logrotate 管理日志
# /etc/logrotate.d/bcachefs

/var/log/bcachefs.log {
    daily
    missingok
    rotate 30
    compress
    delaycompress
    notifempty
    create 0644 root root
}

#!/bin/bash
# backup-snapshot.sh

set -e

MOUNT="/data"
SNAPSHOT_DIR="/data/snapshots"
RETENTION_DAILY=7
RETENTION_WEEKLY=4
RETENTION_MONTHLY=12

TIMESTAMP=$(date +%Y%m%d-%H%M%S)
DAY_OF_WEEK=$(date +%u)
DAY_OF_MONTH=$(date +%d)

# 创建快照
create_snapshot() {
    local type="$1"
    local name="${type}-${TIMESTAMP}"
    echo "创建快照: $name"
    sudo bcachefs subvolume snapshot -r "$MOUNT" "$SNAPSHOT_DIR/$type/$name"
}

# 清理旧快照
cleanup_old() {
    local type="$1"
    local retention="$2"
    
    local count=$(ls "$SNAPSHOT_DIR/$type/" 2>/dev/null | wc -l)
    if [ "$count" -gt "$retention" ]; then
        ls -dt "$SNAPSHOT_DIR/$type/"* | tail -n +$((retention + 1)) | while read snap; do
            echo "删除旧快照: $snap"
            sudo bcachefs subvolume delete "$snap"
        done
    fi
}

# 执行
echo "=== 备份快照 $(date) ==="

# 每日快照
create_snapshot "daily"
cleanup_old "daily" "$RETENTION_DAILY"

# 每周快照 (周日)
if [ "$DAY_OF_WEEK" = "7" ]; then
    create_snapshot "weekly"
    cleanup_old "weekly" "$RETENTION_WEEKLY"
fi

# 每月快照 (1 号)
if [ "$DAY_OF_MONTH" = "01" ]; then
    create_snapshot "monthly"
    cleanup_old "monthly" "$RETENTION_MONTHLY"
fi

echo "=== 备份完成 ==="

#!/bin/bash
# backup-remote.sh

set -e

# 配置
LOCAL_DATA="/data"
REMOTE_HOST="backup-server"
REMOTE_PATH="/backup/data"
REMOTE_USER="backup"
SSH_KEY="/root/.ssh/backup_key"
LOG_FILE="/var/log/backup-remote.log"

echo "=== 远程备份 $(date) ===" | tee -a "$LOG_FILE"

# 使用 rsync 增量同步
rsync -avz --delete \
    -e "ssh -i $SSH_KEY" \
    --exclude='snapshots/' \
    --exclude='tmp/' \
    "$LOCAL_DATA/" \
    "${REMOTE_USER}@${REMOTE_HOST}:${REMOTE_PATH}/" \
    2>&1 | tee -a "$LOG_FILE"

echo "=== 备份完成 $(date) ===" | tee -a "$LOG_FILE"

#!/bin/bash
# backup-borg.sh

set -e

# 配置
export BORG_REPO="ssh://backup@backup-server/backup/borg"
export BORG_PASSPHRASE="your-passphrase"

ARCHIVE_NAME="data-$(date +%Y%m%d-%H%M%S)"

# 初始化仓库（首次）
# borg init --encryption=repokey $BORG_REPO

# 创建备份
borg create \
    --verbose \
    --stats \
    --compression zstd,3 \
    --exclude '/data/snapshots' \
    --exclude '/data/tmp' \
    "::${ARCHIVE_NAME}" \
    /data

# 清理旧备份
borg prune \
    --keep-daily=7 \
    --keep-weekly=4 \
    --keep-monthly=12

echo "备份完成: $ARCHIVE_NAME"

# 生产环境推荐加密
# 创建加密文件系统
sudo bcachefs format --encrypted /dev/sdb

# 使用内核密钥环管理密钥（更安全）
echo "your-secure-passphrase" | keyctl padd user bcachefs-key "" @u

# 挂载时使用密钥环
sudo mount -t bcachefs --keyloc=keyring /dev/sdb /data

# 设置正确的目录权限
sudo chown root:root /data
sudo chmod 755 /data

# 应用数据目录
sudo chown app:app /data/app
sudo chmod 750 /data/app

# 数据库目录
sudo chown postgres:postgres /data/database
sudo chmod 700 /data/database

# 使用 ACL 细化权限
sudo setfacl -R -m g:developers:rwx /data/app
sudo setfacl -R -m g:readonly:rx /data/documents

# 如果使用 SELinux
# 设置正确的上下文
sudo semanage fcontext -a -t user_home_t "/data/app(/.*)?"
sudo restorecon -R /data/app

# 如果使用 AppArmor
# 创建自定义配置文件
cat << 'EOF' | sudo tee /etc/apparmor.d/local/bcachefs-data
/data/** rwk,
EOF

# 每周维护脚本
#!/bin/bash
# weekly-maintenance.sh

set -e

MOUNT="/data"

echo "=== 每周维护 $(date) ==="

# 1. 文件系统检查（dry-run）
echo "文件系统检查..."
sudo bcachefs fsck --dry-run /dev/sdb

# 2. 空间整理
echo "触发 GC..."
sudo bcachefs fs gc "$MOUNT"

# 3. TRIM (SSD)
echo "TRIM..."
sudo fstrim "$MOUNT"

# 4. 清理临时文件
echo "清理临时文件..."
find "$MOUNT/tmp" -type f -mtime +7 -delete

# 5. 检查日志
echo "检查内核日志..."
dmesg | grep -i "bcachefs.*error\|bcachefs.*corrupt" | tail -10

echo "=== 维护完成 ==="

#!/bin/bash
# production-benchmark.sh

set -e

MOUNT="/data"
TEST_DIR="$MOUNT/.benchmark"
RESULT_FILE="/var/log/bcachefs-benchmark-$(date +%Y%m%d).txt"

mkdir -p "$TEST_DIR"

echo "=== 生产环境基准测试 ===" | tee "$RESULT_FILE"
echo "时间: $(date)" | tee -a "$RESULT_FILE"
echo "设备: $(mount | grep $MOUNT | awk '{print $1}')" | tee -a "$RESULT_FILE"
echo "" | tee -a "$RESULT_FILE"

# 清理缓存
sync
echo 3 | sudo tee /proc/sys/vm/drop_caches > /dev/null

# 测试 1: 顺序写入
echo "--- 顺序写入 ---" | tee -a "$RESULT_FILE"
dd if=/dev/zero of="$TEST_DIR/seq-write" bs=1M count=2048 conv=fdatasync 2>&1 | tail -1 | tee -a "$RESULT_FILE"

# 测试 2: 顺序读取
echo "--- 顺序读取 ---" | tee -a "$RESULT_FILE"
echo 3 | sudo tee /proc/sys/vm/drop_caches > /dev/null
dd if="$TEST_DIR/seq-write" of=/dev/null bs=1M 2>&1 | tail -1 | tee -a "$RESULT_FILE"

# 测试 3: 随机 I/O (fio)
if command -v fio &>/dev/null; then
    echo "--- 随机读取 (4K) ---" | tee -a "$RESULT_FILE"
    fio --name=rand-read --ioengine=libaio --rw=randread --bs=4k \
        --size=512M --numjobs=1 --runtime=30 --directory="$TEST_DIR" \
        --output-format=terse 2>/dev/null | awk -F';' '{print "IOPS: "$8", BW: "$7" KB/s"}' | tee -a "$RESULT_FILE"

    echo "--- 随机写入 (4K) ---" | tee -a "$RESULT_FILE"
    fio --name=rand-write --ioengine=libaio --rw=randwrite --bs=4k \
        --size=512M --numjobs=1 --runtime=30 --directory="$TEST_DIR" \
        --output-format=terse 2>/dev/null | awk -F';' '{print "IOPS: "$8", BW: "$7" KB/s"}' | tee -a "$RESULT_FILE"
fi

# 清理
rm -rf "$TEST_DIR"

echo "" | tee -a "$RESULT_FILE"
echo "=== 测试完成 ===" | tee -a "$RESULT_FILE"

#!/bin/bash
# lsyncd-bcachefs.conf

# /etc/lsyncd/lsyncd.conf.lua

settings {
    logfile    = "/var/log/lsyncd.log",
    statusFile = "/var/log/lsyncd.status",
    nodaemon   = false,
}

sync {
    default.rsyncssh,
    source = "/data",
    host = "backup-server",
    targetdir = "/backup/data",
    rsync = {
        archive = true,
        compress = true,
        verbose = true,
    },
    delay = 15,
}

#!/bin/bash
# daily-check.sh

echo "=== Bcachefs 日常检查 $(date) ==="
echo ""

# 1. 文件系统状态
echo "1. 文件系统状态"
mount | grep bcachefs
echo ""

# 2. 空间使用
echo "2. 空间使用"
df -hT | grep bcachefs
echo ""

# 3. 设备状态
echo "3. 设备状态"
for mount in $(mount | grep bcachefs | awk '{print $3}'); do
    echo "  $mount:"
    sudo bcachefs device list "$mount" 2>/dev/null || echo "  无法获取"
done
echo ""

# 4. 内核日志
echo "4. 最近错误"
dmesg | grep -i "bcachefs.*error\|bcachefs.*corrupt" | tail -5 || echo "  无错误"
echo ""

# 5. 快照状态
echo "5. 快照数量"
ls /data/snapshots/daily/ 2>/dev/null | wc -l
ls /data/snapshots/weekly/ 2>/dev/null | wc -l
ls /data/snapshots/monthly/ 2>/dev/null | wc -l
echo ""

echo "=== 检查完成 ==="