# 启用 PG 自动伸缩（强烈推荐）
ceph mgr module enable pg_autoscaler

# 所有池启用自动伸缩
for pool in $(ceph osd pool ls); do
    ceph osd pool set $pool pg_autoscale_mode on
done

# 手动计算参考
# 总 PG 数 ≈ OSD 数 × 100 / 副本数
# 例: 30 OSD，3 副本 → 30 × 100 / 3 = 1000 → 取 1024

#!/bin/bash
# ceph_daily_check.sh - 每日巡检脚本

echo "=================== Ceph Daily Report $(date) ===================="

echo "--- Cluster Status ---"
ceph -s 2>&1

echo ""
echo "--- OSD Status ---"
ceph osd stat
ceph osd tree | grep -E "down|out"

echo ""
echo "--- Capacity ---"
ceph df

echo ""
echo "--- PG Status ---"
ceph pg stat
ceph pg dump_stuck unclean 2>/dev/null | wc -l
ceph pg dump_stuck inactive 2>/dev/null | wc -l
ceph pg dump_stuck stale 2>/dev/null | wc -l

echo ""
echo "--- Slow OSDs ---"
ceph osd perf | sort -k3 -rn | head -5

echo ""
echo "--- Crashes ---"
ceph crash ls-new 2>/dev/null | wc -l

echo ""
echo "--- MON Quorum ---"
ceph mon stat

echo ""
echo "--- MDS Status ---"
ceph mds stat 2>/dev/null

echo ""
echo "--- RGW Status ---"
ceph orch ps | grep rgw

echo ""
echo "=================== End of Report ===================="

# 查看当前生效的配置
ceph config dump

# 查看特定组件的配置
ceph config show osd.0
ceph config show-with-defaults osd.0

# 查看配置差异
ceph config diff

# 配置修改记录
# 所有配置修改应记录在变更管理系统中
# 生产环境修改前先在测试环境验证

#!/bin/bash
# ceph_backup.sh - Ceph 配置和密钥备份

BACKUP_DIR="/backup/ceph/$(date +%Y%m%d)"
mkdir -p $BACKUP_DIR

# 1. 备份 CRUSH Map
ceph osd getcrushmap -o $BACKUP_DIR/crushmap.bin
crushtool -d $BACKUP_DIR/crushmap.bin -o $BACKUP_DIR/crushmap.txt

# 2. 备份所有认证密钥
ceph auth export -o $BACKUP_DIR/ceph.client.admin.keyring
for user in $(ceph auth ls 2>/dev/null | grep -oP 'client\.\w+'); do
    ceph auth get $user -o $BACKUP_DIR/ceph.${user}.keyring
done

# 3. 备份配置文件
cp /etc/ceph/ceph.conf $BACKUP_DIR/ceph.conf

# 4. 备份 Monitor 数据（在其中一个 MON 上执行）
ceph mon getmap -o $BACKUP_DIR/monmap

# 5. 备份 OSD 标记信息
ceph osd dump -o $BACKUP_DIR/osd_dump.txt
ceph osd tree -o $BACKUP_DIR/osd_tree.txt

# 6. 压缩
tar czf /backup/ceph/ceph-backup-$(date +%Y%m%d).tar.gz -C /backup/ceph $(date +%Y%m%d)
rm -rf $BACKUP_DIR

echo "Backup completed: /backup/ceph/ceph-backup-$(date +%Y%m%d).tar.gz"

#!/bin/bash
# rbd_snapshot_backup.sh

POOL="rbd"
IMAGE="myimage"
SNAP="backup-$(date +%Y%m%d)"
BACKUP_DIR="/backup/rbd"

# 创建快照
rbd snap create ${POOL}/${IMAGE}@${SNAP}

# 导出增量备份（基于上一次快照）
if [ -f ${BACKUP_DIR}/${IMAGE}.last_snap ]; then
    LAST_SNAP=$(cat ${BACKUP_DIR}/${IMAGE}.last_snap)
    rbd export-diff ${POOL}/${IMAGE}@${SNAP} \
        --from-snap ${LAST_SNAP} \
        ${BACKUP_DIR}/${IMAGE}-${SNAP}.diff
else
    # 首次全量备份
    rbd export ${POOL}/${IMAGE}@${SNAP} ${BACKUP_DIR}/${IMAGE}-${SNAP}.img
fi

# 记录当前快照
echo ${SNAP} > ${BACKUP_DIR}/${IMAGE}.last_snap

# 1. 确认集群健康
ceph -s  # 必须是 HEALTH_OK

# 2. 确认所有 PG 都是 active+clean
ceph pg stat

# 3. 检查当前版本
ceph versions

# 4. 阅读发行说明
# https://docs.ceph.com/en/latest/releases/

# 5. 设置 noout（防止升级期间触发数据迁移）
ceph osd set noout

# 1. 更新 cephadm
sudo apt update && sudo apt install -y cephadm

# 2. 查看可用版本
ceph orch upgrade check quay.io/ceph/ceph:v18.2.4

# 3. 执行升级
ceph orch upgrade start --ceph-version 18.2.4

# 4. 监控升级进度
ceph orch upgrade status

# 5. 升级完成后验证
ceph -s
ceph versions

# 6. 取消 noout
ceph osd unset noout

# Prometheus 告警规则
# 容量 70% 告警（黄色）
# 容量 85% 告警（红色）
# OSD 使用率偏差过大
# PG 分布不均匀

# 自动扩容脚本（示例）
#!/bin/bash
# auto_scale_check.sh

USAGE=$(ceph df --format json | jq '.stats.total_used_bytes / .stats.total_bytes * 100 | floor')

if [ $USAGE -gt 70 ]; then
    echo "ALERT: Ceph capacity at ${USAGE}%"
    # 发送告警通知
    # 触发扩容流程...
fi