zfs-snapshot-backup/zfs-snapshot-backup.py

import pathlib
import subprocess

SSH_KEY_FILE = "/mnt/Datenspeicher/snap-backup-dataset/backup_key"
SSH_REMOTE = "zfs_snap_backup@etha.cpi.imtek.uni-freiburg.de"

REMOTE_PATH = "zfs-backups"
SCP_REMOTE_URL = f"{SSH_REMOTE}:~/{REMOTE_PATH}/"

ZFS_POOL = "Datenspeicher"
ZFS_ELAB_PREFIX = "elabfs-"

TMP_BACKUP_FOLDER = "/mnt/Datenspeicher/snap-backup-dataset/temporary-backups"


def call(arguments):
    result = subprocess.run(
        arguments, check=True, capture_output=True, text=True
    )
    return result.stdout


def remote_call(arguments):
    cmd = ["ssh", "-i", SSH_KEY_FILE, SSH_REMOTE]
    cmd.extend(arguments)
    return call(cmd)


def clean_split(text):
    items = (item.strip() for item in text.split)
    return [item for item in items if item]


def list_snapshots():
    cmd = ["zfs", "list", "-t", "snapshot", "-H", "-o", "name", "-r", ZFS_POOL]
    return clean_split(call(cmd))


def elab_snapshots():
    result = {}
    for snapshot in list_snapshots():
        pool, snap_name = snapshot.split("/", 1)
        if snap_name.startswith(ZFS_ELAB_PREFIX):
            prefix_and_member, _ = snap_name.rsplit("@", 1)
            member = prefix_and_member.replace(ZFS_ELAB_PREFIX, "")
            if member not in result:
                result[member] = []
            result[member].append(snapshot)
    return result


def gzip_filename(name):
    return f"{name}.gz"


def gzip_filepath(name):
    return pathlib.Path(TMP_BACKUP_FOLDER) / gzip_filename(name)


def list_remote_backups(members):
    result = {}
    for member in members:
        remote_sub_dir = f"{REMOTE_PATH}/{member}"
        try:
            backups = clean_split(remote_call(["ls", remote_sub_dir]))
            result[member] = set(backups)
        except CalledProcessError:
            remote_call(["mkdir", remote_sub_dir])
            result[member] = set()


def backup_latest_snapshot(member, elab_snapshots, existing_backups):
    print(f"backing up member {member}")
    snapshots = sorted(elab_snapshots, reverse=True)
    current_snapshot = snapshots[0]
    latest_backup = None
    for snapshot in snapshots:
        if gzip_filename(snapshot) in existing_backups:
            latest_backup = snapshot
            break
    if current_snapshot == latest_backup:
        # nothing to back up
        print("- nothing to backup, latest snapshot: {current_snapshot}")
        return
    elif latest_backup is None:
        # no snapshot was found in backups, make a full backup for consistency
        send_cmd = ["zfs", "send", current_snapshot]
        print(" - full backup, latest snapshot: {current_snapshot}")
    else:
        # make an incremental backup
        print(
            " - incremental backup, from: {latest_backup} to: {current_snapshot}"
        )
        send_cmd = ["zfs", "send", "-I", latest_backup, current_snapshot]

    # create the backup
    print(" - generating temporary backup file")
    tmp_gzip_filepath = gzip_filepath(current_snapshot)
    with open(tmp_gzip_filepath, "wb") as file_handle:
        gzip_in = subprocess.Popen(
            "gzip", stdin=subprocess.PIPE, stdout=file_handle, check=True
        ).stdin
        subprocess.call(send_cmd, stdout=gzip_in, check=True)

    # copy the backup to the remote server
    print(" - copying temporary backup file")
    remote_url = (
        f"{SSH_REMOTE}:~/{REMOTE_PATH}/{member}/{tmp_gzip_filepath.name}"
    )
    copy_cmd = ["scp", "-i", SSH_KEY_FILE, str(tmp_gzip_filepath), remote_url]
    call(copy_cmd)

    # remove the temporary file
    print(" - removing temporary backup file")
    tmp_gzip_filepath.unlink()


def create_backups():
    elab_snapshots = list_snapshots()
    existing_backups = list_remote_backups(elab_snapshots.keys())
    for member, snapshots in elab_snapshots.items():
        members_backups = existing_backups.get(member, [])
        backup_latest_snapshot(member, snapshots, members_backups)


if __name__ == "__main__":
    create_backups()