You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by payalsharma <pa...@yahoo.com> on 2009/04/21 17:29:39 UTC

Master Slave Solr Replication Automation

We have a requirement of replicating data from one Solr set on a Linux Box to
Second Solr on another Linux box. In order to achieve the same we will use
the SolrCollectionDistributionScripts(snapshooter, snappuller etc) and rsync
utility. 

Configurations:
1.	Apache Solr 1.3.0
2.	Machines : Linux 
3.	Master Slave : 1 Master and 1 slave

Settings done at our end:

Solr on the both Linux boxes contains multiple cores. We have disintegrated
the data to be indexed among multicores, sample solr path of data folder for
the same is like:

Path :: {SOLR_HOME}/solr/multicore/<multi_corename>/data

Sample :
<machine_path>/apache-solr-1.3.0/example/solr/multicore/CORE_WWW.ABCD.COM/data

* SOLR_HOME :: <machine_path>/apache-solr-1.3.0/example/
** multi_corename :: CORE_WWW.ABCD.COM

Thus we will be going to have multiple cores on master as well as slaves
servers

As mentioned on  http:// http://wiki.apache.org/solr/CollectionDistribution 
:For the Solr distribution scripts, the name of the index directory can be
defined by the environment variable data_dir in the configuration file
conf/scripts.conf

	Example conf/scripts.conf file on  slave solr server :
user=
solr_hostname=localhost
solr_port=8080
rsyncd_port=18983
data_dir=${SOLR_HOME}/solr/muticore/CORE_WWW.ABCD.COM /data
webapp_name=solr
master_host=10.x.xx.xxx
master_data_dir=${SOLR_HOME}/solr/muticore/CORE_WWW.ABCD.COM/data
master_status_dir=${SOLR_HOME}/solr/muticore/CORE_WWW.ABCD.COM /status

The index directory name mentioned above should match the value used by the
Solr server which is defined in solr/conf/solrconfig.xml. 

Following are few queries:

1. Please confirm whether the tag entry  : <dataDir></datadir> 
In solrconfig.xml should match for the  Slave solr server / master solr
server in accordance to the scripts.conf configuration settings.
	
2. Also let us know whether some specific handling has to be done in case of
using multi cores during replication.

3. Are there any pitfalls in using the solr distribution scripts and rsync
utility.

Please throw some light on the queries.

-- 
View this message in context: http://www.nabble.com/Master-Slave-Solr-Replication-Automation-tp23158672p23158672.html
Sent from the Solr - User mailing list archive at Nabble.com.