[virt-tools-list] e2fsck an encrypted LVM root partition within a virtual machine

[Keith Roberts]

On 29 Jul 2015, at 12:18, Keith Roberts <keith.roberts at ecric.nhs.uk> wrote:

> Hi all.
> 
> I have a virtual machine with an encrypted LVM root partition that has some fsck errors.
> 
> Is there any way I can use virt-rescue to fix the errors on the encrypted LVM partition please?
> 
> I found this post, but it does not mention fixing the LVM partition inside of a virtual machine.
> 
> http://serverfault.com/questions/375090/using-fsck-to-check-and-repair-luks-encrypted-disk
> 
> Kind Regards,
> 
> Keith Roberts

I have made some progress with this now:

The volume I need to fsck:

# lvm lvs
  LV   VG     Attr      LSize Pool Origin Data%  Move Log Copy%  Convert
  root system -wi-ao--- 8.53g                                           
  swap system -wi-ao--- 1.31g                                           

# lvdisplay -v /dev/system/root
    Using logical volume(s) on command line
  --- Logical volume ---
  LV Path                /dev/system/root
  LV Name                root
  VG Name                system
  LV UUID                xyAkG1-24Zu-tP7k-3eEc-Q8mN-d2ke-6QklGK
  LV Write Access        read/write
  LV Creation host, time linux, 2013-11-07 14:38:34 +0000
  LV Status              available
  # open                 1
  LV Size                8.53 GiB
  Current LE             2184
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     1024
  Block device           253:1


>From inside the running VM:
# df -h
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/system-root  8.3G  2.0G  5.9G  25% /
devtmpfs                 362M   32K  362M   1% /dev
tmpfs                    393M     0  393M   0% /dev/shm
tmpfs                    393M   25M  369M   7% /run
tmpfs                    393M     0  393M   0% /sys/fs/cgroup
tmpfs                    393M   25M  369M   7% /var/lock
tmpfs                    393M   25M  369M   7% /var/run
/dev/xvda1               148M   43M   94M  32% /boot


So now I start virt-rescue on the VM domain file:

# ls
xvda
# virt-rescue -a ./xvda
Could not access KVM kernel module: No such file or directory
failed to initialize KVM: No such file or directory
Back to tcg accelerator.
...
[    8.292332] scsi0 : Virtio SCSI HBA
[    8.310917] scsi 0:0:0:0: Direct-Access     QEMU     QEMU HARDDISK    1.6. PQ: 0 ANSI: 5
[    8.327277] sd 0:0:0:0: [sda] 20971520 512-byte logical blocks: (10.7 GB/10.0 GiB)
[    8.330556] sd 0:0:0:0: [sda] Write Protect is off
[    8.331910] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[    8.557825]  sda: sda1 sda2
[    8.611387] sd 0:0:0:0: [sda] Attached SCSI disk
Cannot find device "eth0"
Cannot find device "eth0"
RTNETLINK answers: Network is unreachable
mdadm: No arrays found in config file or automatically
  No volume groups found
  No volume groups found
------------------------------------------------------------

Welcome to virt-rescue, the libguestfs rescue shell.

Note: The contents of / are the rescue appliance.
You have to mount the guest's partitions under /sysroot
before you can examine them. A helper script for that exists:
mount-rootfs-and-do-chroot.sh /dev/sda2

><rescue> e2fsck /dev/sda1
e2fsck 1.42.8 (20-Jun-2013)
/dev/sda1: clean, 88/40000 files, 52691/159744 blocks
><rescue> 
><rescue> mount-rootfs-and-do-chroot.sh /dev/sda2
mount: unknown filesystem type 'crypto_LUKS'
><rescue> 
><rescue> cryptsetup open /dev/sda2 sysroot
Enter passphrase for /dev/sda2: 
[  355.726880] device-mapper: table: 253:0: crypt: unknown target type
[  355.727799] device-mapper: ioctl: error adding target to table
device-mapper: reload ioctl on temporary-cryptsetup-148 failed: Invalid argument
Failed to open temporary keystore device.
device-mapper: remove ioctl on temporary-cryptsetup-148 failed: No such device or address
...
><rescue> 

Any ideas why I cannot open the LUKS encrypted partition please?

Kind Regards,

Keith