Pymongo - tailing oplog [duplicate]

Question

I am trying to implement a pub/sub on mongo's oplog collection. Provided code works, without tailable = True option set (it will return all docs), but as soon as I pass it to the cursor it won't pickup anything (even after making changes in the desired collection).

I am using pymongo 2.7.2

while(True):
    with self.database.connect() as connection:
        cursor = connection['local'].oplog.rs.find(
            {'ns': self.collection},
            await_data = True,
            tailable = True
        )

        cursor.add_option(_QUERY_OPTIONS['oplog_replay'])

        while cursor.alive:
            try:
                doc = cursor.next()

                print doc
            except(AutoReconnect, StopIteration):
                time.sleep(1)

I have tried few solutions, but it still fails as soon as tailable option is added. Oplog is set up properly, since mongo-oplog module from nodejs works as expected.

Possible duplicate (no accepted answer)

Answer 1

You need to query on the 'ts' oplog field, and keep track of the last document you read (through the timestamp) in case the Cursor has to be recreated. Here's an example you can modify to suit your needs:

import time

import pymongo

c = pymongo.MongoClient()
# Uncomment this for master/slave.
# oplog = c.local.oplog['$main']
# Uncomment this for replica sets.
oplog = c.local.oplog.rs
first = oplog.find().sort('$natural', pymongo.DESCENDING).limit(-1).next()
ts = first['ts']

while True:
    cursor = oplog.find({'ts': {'$gt': ts}}, tailable=True, await_data=True)
    # oplogReplay flag - not exposed in the public API
    cursor.add_option(8)
    while cursor.alive:
        for doc in cursor:
            ts = doc['ts']
            # Do something...
        time.sleep(1)