def verify_backup(rootdir, mdata, num_threads=None):
# Results
changed = []
missing = []
unexpected = []
scan_errors = []
# First, we get all the relevant paths
scan_result = scan_backup(rootdir)
scan_errors = scan_result.errors
unexpected = scan_result.ignored
# Build a dictionary with all original metdata
original_metadata = {}
for p in itertools.chain(mdata.files, mdata.symlinks, mdata.directories):
original_metadata[p.name] = p
# Build a dictionary with all current metadata
digest_map = hashdeep.compute_digests(rootdir, scan_result.files,
num_threads)
uid_map = utils.get_uid_name_map()
gid_map = utils.get_gid_name_map()
current_metadata = {}
for f in scan_result.files:
current_metadata[f] = metadata.get_file_metadata(rootdir, f, digest_map,
uid_map, gid_map)
for s in scan_result.symlinks:
current_metadata[s] = metadata.get_symlink_metadata(rootdir, s, uid_map,
gid_map)
for d in scan_result.directories:
current_metadata[d] = metadata.get_directory_metadata(rootdir, d,
uid_map, gid_map)
# Find missing and unexpected files
all_current_paths = set(scan_result.files + scan_result.symlinks
+ scan_result.directories)
all_original_paths = set(original_metadata.keys())
missing = sorted(list(all_original_paths - all_current_paths))
unexpected.extend(list(all_current_paths - all_original_paths))
unexpected.sort()
# Find changed files
for p in all_current_paths.intersection(all_original_paths):
if not lenient_match(current_metadata[p], original_metadata[p]):
changed.append(p)
# Return the final result
return VerificationResult(changed, missing, unexpected, scan_errors)
#!/usr/bin/env python
import hashdeep
import metadata
import os.path
import pattern
import sys
if __name__ == "__main__":
if len(sys.argv) != 4:
print "usage: %s rootdir patterns_file metadata_file" % sys.argv[0]
sys.exit(1)
rootdir = os.path.normpath(sys.argv[1])
patterns_file = open(sys.argv[2])
patterns = pattern.parse_pattern_file(patterns_file)
pathlist = pattern.assemble_paths(rootdir, patterns)
digests = hashdeep.compute_digests(rootdir, pathlist.filenames)
backup_metadata = metadata.get_backup_metadata(rootdir=rootdir,
files=pathlist.filenames,
symlinks=pathlist.symlinks,
directories=pathlist.directories,
digest_map=digests)
with open(sys.argv[3], "w") as metadata_file:
metadata.write_backup_metadata(metadata_file, backup_metadata)
