mcpField
B
block-storage

block-storage

github.com/bagelhole云与 DevOps1.2k328 次安装23 周前更新
代码仓库

关于 block-storage

面向 Linux 系统管理员的存储管理技能，涵盖磁盘、LVM、AWS EBS、文件系统、快照和软件 RAID 的配置与维护。也提供扩容及故障排查指导。

主要功能

  1. 01磁盘分区与文件系统管理
  2. 02LVM 卷配置与在线扩容
  3. 03AWS EBS 卷管理
  4. 04快照创建与恢复
  5. 05软件 RAID 配置
  6. 06存储故障诊断

适用场景

  1. 01为 Linux 服务器添加和格式化磁盘
  2. 02扩展生产环境中的逻辑卷
  3. 03使用快照进行备份或迁移
  4. 04配置 RAID 冗余或性能

安装

复制下面的命令，在终端中运行。

$npx skills add https://github.com/bagelhole/devops-security-agent-skills --skill block-storage

安装后在 Claude Code 中输入/block-storage确认已加载。

详细介绍

块存储

管理块存储卷，包括 LVM、基于云的 EBS、文件系统创建、快照和 RAID 配置。涵盖从配置原始磁盘到在生产环境中扩展卷的完整生命周期。

适用场景

  • 在 Linux 服务器上添加、分区或格式化新磁盘
  • 管理 LVM 逻辑卷，以灵活分配存储
  • 配置并挂载云块存储（AWS EBS）
  • 创建和恢复快照，以用于备份或迁移
  • 配置软件 RAID，以实现冗余或提升性能
  • 在不影响服务的情况下扩展现有卷

前置条件

  • 拥有目标系统的 root 或 sudo 访问权限
  • 执行 LVM 操作时已安装 lvm2 软件包
  • 执行软件 RAID 操作时已安装 mdadm 软件包
  • 已为 EBS 操作配置 AWS CLI
  • 了解工作负载的 I/O 特性（IOPS、吞吐量）

磁盘发现与分区

# List all block devices
lsblk
lsblk -f    # Show filesystem types and mount points

# Show detailed disk information
fdisk -l /dev/sdb

# Identify disk model and health (requires smartmontools)
smartctl -a /dev/sda
smartctl -H /dev/sda    # Quick health check

# Create a GPT partition table and a single partition
parted /dev/sdb mklabel gpt
parted /dev/sdb mkpart primary ext4 0% 100%

# Alternative: use fdisk for MBR partitioning
fdisk /dev/sdb
# n -> new partition, p -> primary, Enter defaults, w -> write

# Inform the kernel of partition table changes
partprobe /dev/sdb

# Wipe filesystem signatures (prepare for LVM or RAID)
wipefs -a /dev/sdb1

文件系统创建与管理

# Create an ext4 filesystem
mkfs.ext4 /dev/sdb1

# Create an ext4 filesystem with label and reserved block tuning
mkfs.ext4 -L appdata -m 1 /dev/sdb1    # 1% reserved blocks (default is 5%)

# Create an XFS filesystem (recommended for large volumes)
mkfs.xfs /dev/sdb1

# Create an XFS filesystem with label
mkfs.xfs -L appdata /dev/sdb1

# Mount the filesystem
mkdir -p /data
mount /dev/sdb1 /data

# Add persistent mount to fstab (use UUID for reliability)
blkid /dev/sdb1    # Get the UUID
echo 'UUID=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx  /data  ext4  defaults,noatime  0  2' >> /etc/fstab

# Mount all entries in fstab
mount -a

# Check and repair a filesystem (unmount first)
umount /data
fsck.ext4 -y /dev/sdb1
xfs_repair /dev/sdb1    # For XFS

# Resize ext4 (can grow online while mounted)
resize2fs /dev/sdb1

# Resize XFS (must be mounted to grow)
xfs_growfs /data

# Check filesystem usage
df -hT

LVM 管理

创建 LVM 堆栈
# Step 1: Create physical volumes
pvcreate /dev/sdb /dev/sdc

# View physical volumes
pvs
pvdisplay /dev/sdb

# Step 2: Create a volume group from physical volumes
vgcreate data_vg /dev/sdb /dev/sdc

# View volume groups
vgs
vgdisplay data_vg

# Step 3: Create logical volumes
# Fixed size
lvcreate -L 100G -n app_lv data_vg

# Use percentage of free space
lvcreate -l 50%FREE -n logs_lv data_vg

# Use all remaining space
lvcreate -l 100%FREE -n backup_lv data_vg

# View logical volumes
lvs
lvdisplay /dev/data_vg/app_lv

# Step 4: Create filesystem and mount
mkfs.ext4 /dev/data_vg/app_lv
mkdir -p /data/app
mount /dev/data_vg/app_lv /data/app

# Add to fstab
echo '/dev/data_vg/app_lv  /data/app  ext4  defaults,noatime  0  2' >> /etc/fstab
扩展卷（在线）
# Extend a logical volume by 20 GB
lvextend -L +20G /dev/data_vg/app_lv

# Extend to fill all free space in the VG
lvextend -l +100%FREE /dev/data_vg/app_lv

# Grow the ext4 filesystem (online, no unmount needed)
resize2fs /dev/data_vg/app_lv

# Grow XFS filesystem (online)
xfs_growfs /data/app

# Combined: extend LV and resize filesystem in one command
lvextend -L +20G --resizefs /dev/data_vg/app_lv
向现有 VG 添加新磁盘
# Add a new physical volume
pvcreate /dev/sdd

# Extend the volume group
vgextend data_vg /dev/sdd

# Now extend any logical volume using the new space
lvextend -l +100%FREE --resizefs /dev/data_vg/app_lv
LVM 快照
# Create a snapshot (requires free space in VG)
lvcreate -L 10G -s -n app_snap /dev/data_vg/app_lv

# Mount the snapshot read-only for backup
mkdir -p /mnt/snapshot
mount -o ro /dev/data_vg/app_snap /mnt/snapshot

# Perform backup from the snapshot
tar czf /backup/app-$(date +%Y%m%d).tar.gz -C /mnt/snapshot .

# Unmount and remove the snapshot when done
umount /mnt/snapshot
lvremove -f /dev/data_vg/app_snap

# Restore from snapshot (reverts LV to snapshot point -- destructive)
lvconvert --merge /dev/data_vg/app_snap
# Note: if the LV is mounted, merge happens at next activation (reboot)
缩减和移除 LVM 组件
# Shrink a logical volume (ext4 only -- XFS cannot shrink)
# MUST unmount first
umount /data/app
e2fsck -f /dev/data_vg/app_lv
resize2fs /dev/data_vg/app_lv 80G
lvreduce -L 80G /dev/data_vg/app_lv
mount /data/app

# Remove a logical volume
umount /data/app
lvremove /dev/data_vg/app_lv

# Remove a disk from a volume group (migrate data off first)
pvmove /dev/sdc            # Migrate extents to other PVs
vgreduce data_vg /dev/sdc  # Remove PV from VG
pvremove /dev/sdc           # Clean PV metadata

AWS EBS 管理

# Create a gp3 volume (general purpose SSD)
aws ec2 create-volume \
  --availability-zone us-east-1a \
  --size 100 \
  --volume-type gp3 \
  --iops 3000 \
  --throughput 125 \
  --tag-specifications 'ResourceType=volume,Tags=[{Key=Name,Value=app-data},{Key=Environment,Value=production}]'

# Create an io2 volume (provisioned IOPS SSD for databases)
aws ec2 create-volume \
  --availability-zone us-east-1a \
  --size 500 \
  --volume-type io2 \
  --iops 10000

# List volumes with filters
aws ec2 describe-volumes \
  --filters "Name=tag:Environment,Values=production" \
  --query 'Volumes[*].{ID:VolumeId,Size:Size,Type:VolumeType,State:State,AZ:AvailabilityZone}' \
  --output table

# Attach a volume to an instance
aws ec2 attach-volume \
  --volume-id vol-0abc123def456789 \
  --instance-id i-0abc123def456789 \
  --device /dev/xvdf

# After attaching, format and mount on the instance
lsblk                               # Identify the new device (e.g., /dev/nvme1n1)
mkfs.ext4 /dev/nvme1n1
mkdir -p /data
mount /dev/nvme1n1 /data

# Modify a volume (resize without detaching -- gp3/io2)
aws ec2 modify-volume \
  --volume-id vol-0abc123def456789 \
  --size 200

# After resize, grow the filesystem on the instance
growpart /dev/nvme1n1 1              # If partitioned
resize2fs /dev/nvme1n1               # ext4
# xfs_growfs /data                   # XFS

# Create a snapshot
aws ec2 create-snapshot \
  --volume-id vol-0abc123def456789 \
  --description "Pre-upgrade snapshot $(date +%Y-%m-%d)" \
  --tag-specifications 'ResourceType=snapshot,Tags=[{Key=Name,Value=pre-upgrade}]'

# List snapshots
aws ec2 describe-snapshots \
  --owner-ids self \
  --query 'Snapshots[*].{ID:SnapshotId,Vol:VolumeId,Size:VolumeSize,Date:StartTime,Desc:Description}' \
  --output table

# Create a volume from a snapshot (for restore or migration)
aws ec2 create-volume \
  --snapshot-id snap-0abc123def456789 \
  --availability-zone us-east-1a \
  --volume-type gp3

# Detach a volume
aws ec2 detach-volume --volume-id vol-0abc123def456789

# Delete a volume (ensure it is detached first)
aws ec2 delete-volume --volume-id vol-0abc123def456789

软件 RAID（mdadm）

# Install mdadm
apt install -y mdadm    # Debian/Ubuntu
dnf install -y mdadm    # RHEL/CentOS

# Create RAID 1 (mirror) with 2 disks
mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/sdb /dev/sdc

# Create RAID 5 (striped with parity) with 3 disks + 1 spare
mdadm --create /dev/md0 --level=5 --raid-devices=3 --spare-devices=1 /dev/sdb /dev/sdc /dev/sdd /dev/sde

# Create RAID 10 (striped mirrors) with 4 disks
mdadm --create /dev/md0 --level=10 --raid-devices=4 /dev/sdb /dev/sdc /dev/sdd /dev/sde

# Check RAID status
cat /proc/mdstat
mdadm --detail /dev/md0

# Save RAID configuration (persists across reboot)
mdadm --detail --scan >> /etc/mdadm/mdadm.conf     # Debian
mdadm --detail --scan >> /etc/mdadm.conf            # RHEL
update-initramfs -u                                  # Debian

# Create filesystem on RAID device
mkfs.ext4 /dev/md0
mkdir -p /data
mount /dev/md0 /data

# Replace a failed disk
mdadm --manage /dev/md0 --fail /dev/sdc
mdadm --manage /dev/md0 --remove /dev/sdc
# Insert new disk, then:
mdadm --manage /dev/md0 --add /dev/sdf

# Monitor rebuild progress
watch cat /proc/mdstat

# RAID level recommendations:
# RAID 1:  2+ disks, mirroring, good for OS / boot drives
# RAID 5:  3+ disks, single parity, good read performance
# RAID 6:  4+ disks, double parity, survives 2 disk failures
# RAID 10: 4+ disks, mirrored stripes, best I/O performance

故障排查

症状诊断命令常见修复方法
未显示磁盘lsblk、dmesg | tail检查物理连接；重新扫描 SCSI 总线
文件系统为只读dmesg | grep error、mount检测到文件系统错误；卸载后运行 fsck
LVM：VG 中没有可用空间vgs、pvs使用 vgextend 添加新的 PV
未显示 EBS 卷在实例上运行 lsblk在 AWS 控制台中检查挂载状态；NVMe 命名方式有所不同
RAID 降级cat /proc/mdstat使用 mdadm --manage --add 更换故障磁盘
无法调整文件系统大小lvs、df -h先扩展 LV，再调整 FS 大小；XFS 需要处于挂载状态
EBS I/O 速度慢iostat -x 2，检查卷类型升级到 gp3/io2，提高 IOPS/吞吐量
快照耗时过长AWS 控制台：快照进度快照采用增量方式；第一个快照耗时最长

相关技能

  • linux-administration -- 磁盘和文件系统基础
  • performance-tuning -- 使用 fio 调优 I/O 调度器和进行基准测试
  • nfs-storage -- 基于块存储构建的网络文件系统
  • backup-recovery -- 基于快照和文件级别的备份策略
  • object-storage -- 用于非结构化数据的替代存储模型

评论

支持 Markdown