ApacheSolr Beta 2 released
Submitted by Robert Douglass on Sun, 04/20/2008 - 22:49.
Today I released Beta 2 of the ApacheSolr module, the module responsible for search on Robshouse.net. There are some bugfixes, some performance improvements, and a usability improvement, but most of all there is a new feature. With this release it is possible to sort search results. If you search on this site you will notice a block inviting you to Sort by, and your options are Title, Author, Type and Date. Author doesn't make much sense on this site since there is only one, but I think the other sorts will be useful to people who want to find things like the first post I wrote here about Drupal.
I'd also like to encourage people who are using the ApacheSolr module to send me links to their search pages and I'll start compiling a showcase.
LocalSolr
Mike O'Connor just directed my attention to LocalSolr, a Solr with location based searching enabled. Anyone want to give this a shot with Drupal?
-Rob
Hi Rob I want to appreciate
Hi Rob I want to appreciate you in developing this great searching module. I've been using it and eying it for about 2 weeks already. I planned to deploy this module in my site, but prior to that I just want to this issues I have to be answered, if you would be pleased.
1. With over 3 Million items in my test site, in 2 days of running cron, only 19% of total items were indexed. How long would it take me to totally indexed my site considering the setup? I just want to know if this module comes with an option for admins to how many items should be indexed every cron run. I still don't receive permission from my client to use poormanscron or automatic cron run.
2. I want to use apachesolr for general search. A search box/field link by default to core search for every drupal site MUST connect directly to search/apachesolr_search. Any idea where and how to do this?
3. Can 2 solr application (start.jar) be run simultaneously? If yes, consider this scenario: If A & B run start.jar simultaneously, then A run cron, ingesting should be exhibited in A's console. Will ingesting be exhibited also in B's end?
4. These issues here: http://drupal.org/node/add/project-issue/apachesolr .. Are these stuffs addressed or included in beta2?
5. Is ApacheSolr Image integration module working? I tried to enable this one time -- the 3 of them -- and after saving configuration, the whole thing was in mess. Can you have complete documentation on this? It might be a big help.
These are so far my queries and requests. I look forward to your timely and meaty response. Again you've done a great job for this and wish you luck for more stuffs you'll create and share.
Thanks
PS -- I posted this at drupal.org originally but can't wait to have a reply. I needed answers for this badly, and hopefully be attended here. :)
The time it takes to index
The time it takes to index large sites is a problem. Doug Green has a module, reindex, which helps solve this by indexing whole sites in batch mode in Drupal 6, but this isn't available to Drupal 5. Most of the time is spent building nodes (Drupal side), not indexing them (Solr side).
For #2, use the patch and modules in Core Searches to get rid of the built-in content and user searches.
For #3 use Tomcat as your application container and follow the instructions for setting it up using JNDI.
#4 Broken link? Issues in the queue that are open haven't been addressed.
#5 The image integration works on this site... but I haven't tested it other sites.
How long would it take me to
How long would it take me to totally indexed my site considering the setup? I just want to know if this module comes with an option for admins to how many items should be indexed every cron run.
That depends on how many nodes you have.
It runs every cron, and you can probably import around 50 nodes a minute, even on low power hosts. This will vary, though. The number of nodes to do per cron run is configurable.
-Rob
Post new comment