Closed GoogleCodeExporter closed 8 years ago
there are two ways I know to setup solr with multiple servers.
The first is the traditional 1 master to 1 or more slaves fanout. In this
setup, replication of the index changes occurs periodically, so there's always
a minimum delay between when changes on the master are committed and when the
slave will have those changes.
The second is to do what's called sharding. Where you split responsibility for
different documents in your index across all the servers. Each search is
responsible for updating and searching its shard of the index - which means
there's not replication delay. When a query happens, you actually tell the
server that it should query and aggregate the results from several shards. With
the 3.x series this required careful setup and maintenance by you (you actually
had to pass the shards to use in your query parameters). In 4.x (still only a
release candidate) they're moving that configuration information into zookeeper
to make the server coordination more automatic. If you really want near real
time search, I imagine that's the way they'll want you to go.
Please look at the solr wiki and mail lists: http://lucene.apache.org/solr
Since this is really a solr server configuration and setup question. Hope I
helped though.
Original comment by donovan....@gmail.com
on 13 Jul 2012 at 1:37
I mean there are actually two servers in the following two server shared drive.
Indexing server, just indexing the work you do. Let the search server, this Is
it possible to search indexed documents at the same time?
Thanks prev reply
Original comment by hasani...@gmail.com
on 13 Jul 2012 at 1:45
I don't think that is a suggested way to do it (shared drive). Each time a
commit happens an index reader needs to be reopened - i don't think you'd see
the changes in the server that didn't do the commit. It wouldn't know something
happened.
You CAN search on the server doing updates though. If that wasn't obvious.
Again, I think this is a better question for the solr user mailing list.
Original comment by donovan....@gmail.com
on 13 Jul 2012 at 1:50
thanks for the helps.
like you said index reader needs to re-open the index. I think the structure
does not happen. I think the solution will be to replication
Original comment by hasani...@gmail.com
on 13 Jul 2012 at 2:03
Original issue reported on code.google.com by
hasani...@gmail.com
on 13 Jul 2012 at 8:14