keeper.py

#!/usr/bin/env python

#
# (c) Krzysztof Magosa
# License: MIT
# Use at your own risk.
#
# https://gist.github.com/krzysztof-magosa/d9d96c8864a20dc3b47fd1eee1cfcc4f
#

import argparse
import os
import heapq

parser = argparse.ArgumentParser()
parser.add_argument("--thresh", type=int, default=80, help="Maximum used space in %%")
parser.add_argument("--offenders", type=int, default=4096, help="Find N offenders at a time")
parser.add_argument("--hysteresis", type=int, default=5, help="Remove N more %%")
parser.add_argument("path", nargs=1, help="Path to scan")
args = parser.parse_args()


# Return % of disk used
def get_disk_percents(path):
    vfs = os.statvfs(path)
    return 100.0 * float((vfs.f_blocks - vfs.f_bfree) / float(vfs.f_blocks - vfs.f_bfree + vfs.f_bavail))


# Return files from path (recursively) as tuple of mtime+filename
def get_files(path):
    for root, dirs, files in os.walk(path):
        for name in files:
            filename = os.path.join(root, name)

            try:
                # Only regular files
                if os.path.isfile(filename):
                    mtime = os.path.getmtime(filename)
                    yield (mtime, filename)
            except:
                pass


# Get offenders from path
# Priority Queue is sorted by first element from tuple (mtime).
# Therefore files having oldest mtime are returned.
def get_offenders(path, limit):
    offenders = []
    for entry in get_files(path):
        heapq.heappush(offenders, entry)

        if len(offenders) > limit:
            del offenders[-1]

    return offenders


if __name__ == "__main__":
    # Initial check whether we should do anything.
    if get_disk_percents(args.path[0]) > args.thresh:
        offenders = []

        while True:
            # Remove to reach thresh + hysteresis.
            if get_disk_percents(args.path[0]) <= (args.thresh - args.hysteresis):
                break

            # We have no offenders available, find them.
            if not offenders:
                offenders = get_offenders(args.path[0], args.offenders)

            # No offenders available, nothing we can do.
            if not offenders:
                break

            # Remove oldest offender ignoring possible permission denied problems.
            mtime, filename = heapq.heappop(offenders)
            try:
                print("Removing {}".format(filename))
                os.unlink(filename)
            except OSError:
                pass
    else:
        print("Usage below threshold. Exiting...")