Sharding and Replication setup

[erolagnab <tr...@gmail.com>]

Hi all,

I'm trying to setup Solr in current environment to provide high availabilty
and fault torrenlance infrastucture.
What I have is:
1. 2 physical servers running 2 Tomcats
2. A load balancer doing round-robin requests to 2 Tomcats

After reading thru different posts
(http://lucidworks.lucidimagination.com/display/solr/Scaling+and+Distribution,
http://www.slideshare.net/sourcesense/sharded-solr-setup-with-master), I'm
thinking of having the setup as in the image attached 
http://lucene.472066.n3.nabble.com/file/n4001642/Solr_Sharding_and_Replication_HA.png
Solr_Sharding_and_Replication_HA.png 

Basically, it's a combination of sharding and replication.

Server 1, I have a 3-core Solr instance. e.g.: Master 1, Slave 1 (replicated
locally from Master 1) and Slave 2 (replicated remotely from Master 2 in
Server 2)
Server 2, I have similar setup, e.g.: Master 2, Slave 2 (replicated locally
from Master 2) and Slave 1 (replicated remotely from Master 1 in Server 1)

The indexing request can come to Master 1 or Master 2 hence Slave 1 and
Slave 2 become high available shards (as in both servers).

The search requests are served by a virtual coordinator (can hit slave 1
core or slave 2 core directly with shards parameter) to combine results from
Slave 1 and Slave 2.

I haven't done the actual implementation yet. Just post here so to hear any
suggestion/recommendation/pitfalls/gotchas on this setup from experts.

Very appreciate your attention.

Cheers,

Trung



--
View this message in context: http://lucene.472066.n3.nabble.com/Sharding-and-Replication-setup-tp4001642.html
Sent from the Solr - User mailing list archive at Nabble.com.