Skip to content

Instantly share code, notes, and snippets.

Embed
What would you like to do?
Recover deleted data from mongo DB database
"""A little script to recover deleted recording of a mongoDB db file
There's no optimization but it work and has saved me
"""
import struct
import bson
import pymongo
import sys
def decode_chunck(chunck):
"Try to decode a chunck"
#if not bson.is_valid(chunck):
# return None
try:
result = bson.decode_all(chunck)[0]
if not result:
return None
else:
# if there's all the searched field, return it
if 'field_1' in result and 'field_2' in result and 'field_3' in result:
return result
except Exception:
return None
return None
def generate_chunck(data, pos=0):
"Generator to create chunck"
print "open at: %s" % pos
f= open(data,'rb')
a=f.read()
size = len(a)
while pos < size:
# Progress indicator
if pos % 1024 ==0:
print pos
# Determine the size of the possible bson encoded data
bson_size = struct.unpack("<I", a[pos:pos + 4])[0]
# If it's more than 2KB reject it (perfect for me)
if bson_size > 2*1024:
# Continue tu search in the file
pos += 1
continue
# If the bson is bigger than the file, reject it
if bson_size+pos > size-1:
pos += 1
continue
# A bson should end by \x00
# http://bsonspec.org/#/specification
if a[pos+bson_size] != '\x00':
pos += 1
continue
# Chunck it
chunck = a[pos:pos+bson_size]
pos += 1
yield chunck
# create connection
connection = pymongo.Connection('localhost', 27017)
# Connect to MongoDB in order to reinsert the data
db = connection.recover_db
collection = db.recover_collection
# argv[1] = the file to recover
# argv[2] = Where to start in the file
for chunck in generate_chunck(sys.argv[1], int(sys.argv[2])):
result = decode_chunck(chunck)
if result:
print "insert"
collection.insert(result)
@sforce100
Copy link

sforce100 commented Jul 5, 2013

hi man, this is to recover some data of one database, but can not recover a drop database, is it?

@magnums
Copy link

magnums commented Apr 26, 2014

Could you please show sample of argv1 and argv2?
and Where can found the file in both argv ?
Thank you in advance.

@guss77
Copy link

guss77 commented Nov 14, 2015

This didn't work for me on Mongo 2.4, apparently the file format has changed and now mongo overwrites the object size with \xee\xee\xee\xee. I have a gist here: https://gist.github.com/guss77/f8e610cfddbe02c07896 that modifies the above code to try to guess at the recoverable size of an object.

@atul9911
Copy link

atul9911 commented Jan 7, 2017

Showing : ./recover.py: line 8: syntax error near unexpected token (' ./recover.py: line 8: def decode_chunck(chunck):

@arihantdaga
Copy link

arihantdaga commented Sep 29, 2018

@egguy can u please give an example for argv[1] and argv[2], since i dont know which file do we have to feed it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment