We are currently running Solr 4.6.1 in a dev/testing environment running on
top of CentOS 6.x 64bit w/12 gigs ram.
Not being real familiar with Linux and Solr we basically just copied the
Solr-4.6.1 folder/file structure right into the "var" folder of the server
and launched the Solr engine with the following command:
java
-Djetty.class.path=/var/solr-4.6.1/example/solr-webapp/webapp/WEB-INF -jar
start.jar
Our production Solr environment is expected to used the same hardware spects
and to be pretty basic as our document number (less that 50k documents) and
our index size (< 60 meg) is rather small. We are looking at a single local
server running a single instance of Solr. Replication is not required as
the time to rebuild the index in case of data lost is minimal. As for fault
tolerance we may run a 2nd server with a identical setup as a fail over if
needed.
In prep for a production setup I have a few questions that I would like to
get some feedback on:
1) By default everything is running from under the "example" folder which
does not make much sense for a production environment. So is it possible to
re-organize the folder structure and remove the "Example" parent folder or
are we better off running it with the folder structure as is?
1.a) If we can reorganize it and remove the "example" folder are there
any config files that need to be modified to support this change?
2) By default we re running with the Jetty web server. Is this ok or should
we use Tomcat as most of the "install" related posts/blogs reference Tomcat
in their instructions?
3) Based on the server specs mentioned above is there any Java related
configs we should implement to make the best use of our resources for the
best overall performance.
4) Anything else we should be aware of/watch out for?
Thank
Lee.
--
View this message in context:
http://lucene.472066.n3.nabble.com/Solr-Production-Installation-tp4122091.html
Sent from the Solr - User mailing list archive at Nabble.com.
