pgfid_index.py

import uuid
import xattr
import os
from hashlib import md5
import sys

IGNORE_DIRS = [".glusterfs", ".trashcan"]
PFX = "user.pgfid"


def crawl(brick_path, path, ignore_dirs=[]):
    if path in ignore_dirs:
        return

    pgfid = str(uuid.UUID(bytes=xattr.get(path, "trusted.gfid",
                                          nofollow=True)))
    # pgfid = str(uuid.UUID(bytes=xattr.get(path, "glusterfs.gfid",
    #                                       nofollow=True)))

    files = os.listdir(path)
    for f in files:
        p = os.path.join(path, f)
        if os.path.isdir(p):
            crawl(brick_path, p, ignore_dirs)
        else:
            hash = md5(f).hexdigest()
            xattr_key = "{pfx}.{pgfid}.{hash}".format(pfx=PFX,
                                                      pgfid=pgfid,
                                                      hash=hash)
            xattr_value = f
            xattr.set(p, xattr_key, xattr_value, nofollow=True)


if __name__ == "__main__":
    brick = sys.argv[1]
    ignore_dirs = [os.path.join(brick, v) for v in IGNORE_DIRS]
    crawl(brick, brick, ignore_dirs)