Controlling read/write access for replicated indexes

[Jeff Rodenburg <je...@gmail.com>]

I'm curious what mechanisms everyone is using to control read/write access
for distributed replicated indexes.  We're moving to a replication
environment very soon, and our client applications (quite a few) all have
configuration pointers to the URLs for solr instances.  As a precaution, I
don't want errant configuration values to inadvertently send write requests
to read servers, as an example.  As an aside, we're running solr under
tomcat 5.5.x which has its own control aspects as well.

Any best practices, i.e. something that's not a maintenance headache later,
from those who have done this would be greatly appreciated.

thanks,
j.r.

Re: Controlling read/write access for replicated indexes

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

Solr's web.xml has a servlet-mapping which maps url-pattern "/update/*" to
the SolrUpdateServlet.  You can set up a <security-constraint> on that to
protect against inadvertent updates.

Bill

On 3/28/07, Jeff Rodenburg <je...@gmail.com> wrote:
>
> I'm curious what mechanisms everyone is using to control read/write access
> for distributed replicated indexes.  We're moving to a replication
> environment very soon, and our client applications (quite a few) all have
> configuration pointers to the URLs for solr instances.  As a precaution, I
> don't want errant configuration values to inadvertently send write
> requests
> to read servers, as an example.  As an aside, we're running solr under
> tomcat 5.5.x which has its own control aspects as well.
>
> Any best practices, i.e. something that's not a maintenance headache
> later,
> from those who have done this would be greatly appreciated.
>
> thanks,
> j.r.
>