snapshot

Snapshot Management

Overview

Skills for LVM snapshots, filesystem snapshots, and cloud snapshot management.

LVM Snapshots

Create Snapshot

# 查看逻辑卷
lvs
lvdisplay

# 创建快照
lvcreate -L 10G -s -n snap_data /dev/vg0/data

# 创建薄快照
lvcreate -s -n snap_data /dev/vg0/thin_data

View Snapshots

lvs -a lvdisplay /dev/vg0/snap_data

### 挂载快照
```bash
# 挂载只读
mount -o ro /dev/vg0/snap_data /mnt/snapshot

# 挂载读写
mount /dev/vg0/snap_data /mnt/snapshot

Restore Data

# 从快照恢复
lvconvert --merge /dev/vg0/snap_data

# 恢复前需卸载原卷
umount /dev/vg0/data
lvconvert --merge /dev/vg0/snap_data
mount /dev/vg0/data /data

Delete Snapshot

# 卸载并删除
umount /mnt/snapshot
lvremove /dev/vg0/snap_data

Btrfs Snapshots

Create Snapshot

# 创建只读快照
btrfs subvolume snapshot -r /data /snapshots/data_$(date +%Y%m%d)

# 创建可写快照
btrfs subvolume snapshot /data /snapshots/data_writable

# 查看子卷
btrfs subvolume list /

Manage Snapshots

# 删除快照
btrfs subvolume delete /snapshots/data_20240101

# 发送快照（备份）
btrfs send /snapshots/data_readonly | btrfs receive /backup/

# 增量发送
btrfs send -p /snapshots/old /snapshots/new | btrfs receive /backup/

ZFS Snapshots

Create Snapshot

# 创建快照
zfs snapshot pool/dataset@snap_$(date +%Y%m%d)

# 递归创建
zfs snapshot -r pool/dataset@snap_name

# 查看快照
zfs list -t snapshot

Manage Snapshots

# 回滚
zfs rollback pool/dataset@snap_name

# 克隆
zfs clone pool/dataset@snap_name pool/clone_dataset

# 删除
zfs destroy pool/dataset@snap_name

# 发送/接收
zfs send pool/dataset@snap | zfs receive backup/dataset

Cloud Snapshots

AWS EBS

# 创建快照
aws ec2 create-snapshot \
    --volume-id vol-1234567890abcdef0 \
    --description "Backup $(date +%Y%m%d)"

# 查看快照
aws ec2 describe-snapshots --owner-ids self

# 从快照创建卷
aws ec2 create-volume \
    --snapshot-id snap-1234567890abcdef0 \
    --availability-zone us-east-1a

# 删除快照
aws ec2 delete-snapshot --snapshot-id snap-1234567890abcdef0

Alibaba Cloud

# 创建快照
aliyun ecs CreateSnapshot --DiskId d-xxx --SnapshotName backup

# 查看快照
aliyun ecs DescribeSnapshots

# 删除快照
aliyun ecs DeleteSnapshot --SnapshotId s-xxx

Common Scenarios

Scenario 1: Database Consistent Snapshot

#!/bin/bash
# MySQL + LVM 快照
mysql -e "FLUSH TABLES WITH READ LOCK;"
lvcreate -L 10G -s -n db_snap /dev/vg0/mysql_data
mysql -e "UNLOCK TABLES;"

# 备份快照
mount -o ro /dev/vg0/db_snap /mnt/snap
tar -czvf /backup/mysql_$(date +%Y%m%d).tar.gz /mnt/snap
umount /mnt/snap
lvremove -f /dev/vg0/db_snap

Scenario 2: Automated Snapshot Script

#!/bin/bash
# Btrfs 自动快照
SNAP_DIR="/snapshots"
MAX_SNAPS=7

# 创建快照
btrfs subvolume snapshot -r /data ${SNAP_DIR}/data_$(date +%Y%m%d_%H%M)

# 清理旧快照
ls -1d ${SNAP_DIR}/data_* | head -n -${MAX_SNAPS} | xargs -r btrfs subvolume delete

Scenario 3: Pre- and Post-Snapshot Hooks

#!/bin/bash
# 快照前
systemctl stop application
sync

# 创建快照
lvcreate -L 5G -s -n app_snap /dev/vg0/app_data

# 快照后
systemctl start application

Troubleshooting

Issue	Troubleshooting Method
Snapshot Space Full	Extend snapshot, reduce changes
Snapshot Invalid	Check COW space
Restore Failed	Check volume status and dependencies
Performance Degradation	Reduce the number of snapshots

# LVM 快照状态
lvs -a -o +snap_percent

# Btrfs 空间
btrfs filesystem df /
btrfs filesystem usage /

# ZFS 空间
zfs list -o space