Elasticsearch river and mongodb replica


(Robbie Cheng) #1

According to the documentation of elasticsearch river to mongodb. It
requires oplog to pull data from mongo. We tired to run a replication
set with single node (yes, cheating) instructed by the following url:

And it looks like river pulls data from mongodb successfully, but we
would like to know more detail since we only spotted 2 records in
oplog, how does river pull existing data from replication set? And is
there any dependency of creating replication set in mongodb, and
hooking up ES and mongodb with river? I mean, if we have running
replication set, and we hook it up with ES afterwards, what happen to
those existing data? Will they be pulled, too?

Thanks,

--


(Richard Louapre) #2

Hi,

You can find more details oplog and internal replication in MongoDB here
[1].
This is how currently the data are pulled from mongo inot ES.

[1] - http://www.kchodorow.com/blog/2010/10/12/replication-internals/

Thanks,
Richard.

On Tuesday, October 23, 2012 1:03:22 PM UTC-4, Robbie Cheng wrote:

According to the documentation of elasticsearch river to mongodb. It
requires oplog to pull data from mongo. We tired to run a replication
set with single node (yes, cheating) instructed by the following url:

http://loosexaml.wordpress.com/2012/09/03/how-to-get-a-mongodb-oplog-without-a-full-replica-set/

And it looks like river pulls data from mongodb successfully, but we
would like to know more detail since we only spotted 2 records in
oplog, how does river pull existing data from replication set? And is
there any dependency of creating replication set in mongodb, and
hooking up ES and mongodb with river? I mean, if we have running
replication set, and we hook it up with ES afterwards, what happen to
those existing data? Will they be pulled, too?

Thanks,

--


(system) #3