collectionDistribution question

[Mochamad bahri nurhabbibi <he...@gmail.com>]

Hello all.. I am new in SOLR.. previously I asked question about searcher,
many thanks to  acidbriggs for answering my mail.

I have a question about collection & distribution from
http://wiki.apache.org/solr/CollectionDistribution page

from the terminology part, about master/slave

The Solr distribution system uses the master/slave model. The master is the
service which receives all updates initially and keeps everything organized.
Solr uses a single update master server coupled with multiple query slave
servers. All changes (such as inserts, updates, deletes, etc.) are made
against the single master server. Changes made on the master are distributed
to all the slave servers which service all query requests from the clients.

from that part, I can conclude :

1. updates to data made only to master server
2. updates will be distributed to slave
3. search queries from clients are distributed to slave.

questions :

1. is my conclusion right ? please correct me if I am wrong
2. when master server optimizing the index what will happen if someone
updates the data  ? do they have to wait or something else happen ?
3. do I need to replicate the master server ? what is the change performance
of this server ?

thank you

-regards-
habibi

Re: collectionDistribution question

[Mochamad bahri nurhabbibi <he...@gmail.com>]

thanks for your answer & illustration Bill..

I guess that :

during the optimization, the update request will be recorded into one file,
then after the optimization, it will be merged.

Do u think that I should replicate the master ? to make update can perform
seamlessly ?


One more thing.. about the replication strategy...

1. What is the drawback of using 1 tomcat with 3 index with one solr.war ?
2. Is there burden to the core if we use 3 index ?
3. which one is better if we use 3 solr.war each with 1 index ? (installed
upon 3 different apache)

thank you.

Re: collectionDistribution question

[Bill Au <bi...@gmail.com>]

Your first and second conclusions are right, but not the third one.

Solr does not distribute the queries from clients to the slaves.  You need
to do that yourself (for example, use a load balancer in front of your pool
of Solr slaves).

I think that when Solr is optimizing the index all update requests will be
blocked until the optimize is done.

The way collection distribution is set up you can have only one master.  But
you  can chain them though ( ie machine A distribute to machine B, machine B
then in turn distribute to machine C and machine D.

Bill



On Feb 13, 2008 5:36 AM, Mochamad bahri nurhabbibi <he...@gmail.com>
wrote:

> Hello all.. I am new in SOLR.. previously I asked question about searcher,
> many thanks to  acidbriggs for answering my mail.
>
> I have a question about collection & distribution from
> http://wiki.apache.org/solr/CollectionDistribution page
>
> from the terminology part, about master/slave
>
> The Solr distribution system uses the master/slave model. The master is
> the
> service which receives all updates initially and keeps everything
> organized.
> Solr uses a single update master server coupled with multiple query slave
> servers. All changes (such as inserts, updates, deletes, etc.) are made
> against the single master server. Changes made on the master are
> distributed
> to all the slave servers which service all query requests from the
> clients.
>
> from that part, I can conclude :
>
> 1. updates to data made only to master server
> 2. updates will be distributed to slave
> 3. search queries from clients are distributed to slave.
>
> questions :
>
> 1. is my conclusion right ? please correct me if I am wrong
> 2. when master server optimizing the index what will happen if someone
> updates the data  ? do they have to wait or something else happen ?
> 3. do I need to replicate the master server ? what is the change
> performance
> of this server ?
>
> thank you
>
> -regards-
> habibi
>