Return a volume to a consistent snapshot

Situation

This machine has a volume group lab_vg. The data volume is mounted at /mnt/lvm-data, the lab-writer service works with it, and next to it sits the keep volume with other data. After a training update, /mnt/lvm-data/settings.conf says mode=broken and a new file after.txt appeared. A classic snapshot /dev/lab_vg/checkpoint was taken before the update. Everything that changed on data after the snapshot is to be discarded.

Task

Roll the data volume back to the snapshot and put it back in service:

1. /mnt/lvm-data/settings.conf says mode=stable again, after.txt is gone, and the rest of the data (data/accounts.csv, owners and modes) is as it was before the update;
2. the original data volume is mounted read-write at /mnt/lvm-data: the same logical volume and the same filesystem, 128 MiB in size;
3. no separate checkpoint snapshot and no unfinished merge remain;
4. the lab-writer service is running again, now on the restored volume;
5. the keep volume, the volume group, /etc/fstab and other mounts are untouched.

Constraints

• Leave keep alone: don't mount it or change it.
• Don't create new volumes or filesystems, and don't edit /etc/fstab.

The tools you need are already installed. There's no internet on this machine.