Disk Issues Lab

The lab in this module must be run on server1.example.com.

Goal

Recover corrupted LVM metadata

1. LVM Recovery

Look at a list of current VGs, then create a backup of the LVM metadata for the extra volume.
```
# vgs
# vgcfgbackup extra
# ls -al /etc/lvm/backup
```
Create a VM snapshot if your hypervisor supports it.
Simulate corruption of the LVM metadata on the extra disk, then reboot the system (never do this in practice!).
```
# dd if=/dev/zero of=/dev/vdb1 count=10
```
Check for the extra VG; it should no longer appear.
```
# vgs
```

Find the PV’s UUID and device name from /etc/lvm/backup/extra:

 physical_volumes {
                pv0 {
                       id = "x8gEhX-zLGb-GUJg-MkqT-7wlr-FH3v-use23e"
                       device = "/dev/vdb1"    # Hint only
                       status = ["ALLOCATABLE"]
                       flags = []
                       dev_size = 39483392     # 18.8271 Gigabytes
                       pe_start = 2048
                       pe_count = 602  # 18.8125 Gigabytes
                }
        }

Before restoring the VG, you must reboot the system since the VG was removed uncleanly.
```
# reboot
```

Repair the PV’s metadata (ignore the error about not being able to find the device).

# pvcreate --uuid <UUID FOR PV FOUND IN BACKUP> \
    --restorefile /etc/lvm/backup/extra /dev/<DEVICE FOUND IN BACKUP>

Restore the VG, re-enable the VG, and then verify.

# vgcfgrestore extra
# vgchange -a y extra
# vgs
# mount /dev/extra/disk /mnt/temp
# ls /mnt/temp
# md5sum /etc/passwd /mnt/temp/passwd