[sheepdog-users] Correct vdi restore procedure

[Valerio Pachera]

Do you confirm the following procedure?

Consider vdi 'wheezy' with snapshots 1,2

Snapshot 1 is the older one.

I have a full copy of snapshot 1

  qemu-img convert -O qcow2 sheepdog:wheezy:1 /media/nas/whezy_s1.qcow2

I save all the increments

  dog vdi backup -F 1 -s 2 wheezy > /media/nas/inc2.img

Considering a disater recovery

  dog cluster format
  qemu-img convert -f qcow2 /media/nas/whezy_s1.qcow2 sheepdog:wheezy
  dog vdi snapshot wheezy
  dog vdi list
    s wheezy     1   10 GB  1.8 GB  0.0 MB 2013-08-30 17:58   5ddf88
  2
    wheezy     0   10 GB  0.0 MB  1.8 GB 2013-08-30 18:01   5ddf9a     2
  dog vdi restore -s 1 wheezy < /media/nas/inc2.img
  dog vdi list
    s wheezy_template     1   10 GB  1.8 GB  0.0 MB 2013-08-30 17:58
5ddf88     2
    s wheezy_template     2   10 GB  1.0 GB 1008 MB 2013-08-30 18:02
5ddf89     2
      wheezy_template     0   10 GB   96 MB  1.7 GB 2013-08-30 18:02
5ddf9b     2
  # If we start the guest right now it would start as of snapshot 1,
so we roll back (ahead in this case: )
  dog vdi rollback -s 2 wheezy

In case of more increments, repeat the necessary steps.