问题描述
虽然需要恢复数据，但并非应 LVM 经常故障，而是人为原因导致数据丢失。
该笔记将记录：在 Linux 中，如何从 LVM 中恢复数据，以及常见故障场景的处理方法。
解决方案
我们并非数据恢复专家，很多内容都参考自网络：
1）Recovering a Lost LVM Volume Disk
场景一、物理卷被格式化，且无元数据备份
将显示 unknown device 信息：
# pvs WARNING: Couldn't find device with uuid A11H2H-99zO-Urtx-NUkn-uRE3-umXU-P1hIWj. WARNING: VG dt is missing PV A11H2H-99zO-Urtx-NUkn-uRE3-umXU-P1hIWj (last written to /dev/nvme0n1p4). PV VG Fmt Attr PSize PFree /dev/sda3 ubuntu-vg lvm2 a-- <9.00g 0 /dev/sdb10 dt lvm2 a-- 48.82g 0 /dev/sdb11 dt lvm2 a-- 48.82g 0 /dev/sdb12 dt lvm2 a-- 39.32g <9.94g /dev/sdb5 dt lvm2 a-- 48.82g <1.18g /dev/sdb6 dt lvm2 a-- 48.82g 0 /dev/sdb7 dt lvm2 a-- 48.82g 0 /dev/sdb8 dt lvm2 a-- 48.82g 0 /dev/sdb9 dt lvm2 a-- 48.82g 48.82g [unknown] dt lvm2 a-m 48.82g 0
结合 lsblk 输出：
# lsblk /dev/sdb NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT sdb 8:16 0 477G 0 disk ├─sdb1 8:17 0 300M 0 part ├─sdb2 8:18 0 128M 0 part ├─sdb3 8:19 0 46.6G 0 part /mnt/fix/rootfs ├─sdb4 8:20 0 48.8G 0 part ├─sdb5 8:21 0 48.8G 0 part ├─sdb6 8:22 0 48.8G 0 part ├─sdb7 8:23 0 48.8G 0 part ├─sdb8 8:24 0 48.8G 0 part ├─sdb9 8:25 0 48.8G 0 part ├─sdb10 8:26 0 48.8G 0 part ├─sdb11 8:27 0 48.8G 0 part └─sdb12 8:28 0 39.3G 0 part
我们得知磁盘 /dev/sdb4 被格式化，因此使用它重新创建 PV 实例
好在 LVM 以纯文本形式存储相当大的头信息，包括多数据的多个备份：
# strings /dev/sdb5 | less // 查看这些信息 ... # vgcfgbackup dt // 所以虽然 PV 损坏，但依旧能够完成备份； WARNING: PV /dev/sdb4 in VG dt is missing the used flag in PV header. Volume group "dt" successfully backed up. # pvcreate --uuid A11H2H-99zO-Urtx-NUkn-uRE3-umXU-P1hIWj -ff --restorefile /etc/lvm/backup/dt /dev/sdb4 WARNING: Couldn't find device with uuid A11H2H-99zO-Urtx-NUkn-uRE3-umXU-P1hIWj. WARNING: Couldn't find device with uuid ARQRI6-whVg-o2B2-O1qK-rh8c-Qdo0-XmhWjO. WARNING: Couldn't find device with uuid s5FrBC-QhyM-h7DR-qzvE-F6Wj-JiG0-zEK1R4. WARNING: Couldn't find device with uuid XHt33t-lOEO-mzQ9-XT66-sWdo-S0HD-NWjkQA. WARNING: Couldn't find device with uuid FIAWwN-t3cx-mbgA-zdVX-ayJ3-Xfu7-1dJJtj. WARNING: Couldn't find device with uuid JlkM1S-LcL2-g6VO-3e30-j6gM-z82f-cUNELP. WARNING: Couldn't find device with uuid DLdq59-rQ1w-3jow-U1Nz-CcZL-3oPG-q3IJ92. WARNING: Couldn't find device with uuid 8s2vFL-gJ3c-kbcD-kAHR-l1vC-9eUQ-neRSWE. WARNING: Couldn't find device with uuid oiBwX0-GFxZ-qApn-ZrcI-ARPN-rDGU-3cGwRY. WARNING: PV /dev/sdb4 in VG dt is missing the used flag in PV header. Really INITIALIZE physical volume "/dev/sdb4" of volume group "dt" [y/n]? y WARNING: Forcing physical volume creation on /dev/sdb4 of volume group "dt" WARNING: PV /dev/sdb4 in VG dt is missing the used flag in PV header. Physical volume "/dev/sdb4" successfully created. # vgcfgrestore dt Restored volume group dt. # vgscan Found volume group "dt" using metadata type lvm2 Found volume group "ubuntu-vg" using metadata type lvm2 # vgchange -ay dt 4 logical volume(s) in volume group "dt" now active
文件系统修复：
# e2fsck -y /dev/dt/opt ...
至此已经完成大部分工作。但是：
1）由于磁盘已经格式化，所以当命令 e2fsck 执行结束后，依旧会丢失部分数据；
2）如果被格式化的分区是文件系统最开始的部分，即元数据丢失，则只能借助其他工具恢复文件；
参考文献
Recovering a Lost LVM Volume Disk
can you recover lvm metadata without access to /etc/lvm/?
How to recover deleted Logical volume (LV) in LVM using vgcfgrestore