Solr is slow when indexing

Question

I have over 100 CSV files that have 10,000 lines that I index. And then the spelling request is like spelling. At the same time, indexing is very slow.

I found some good solutions

• Master slave, where the master pointer and slave are used for the request. How to index records in Solr faster (and not affect the ColdFusion web server)? Two JVMs?

• Using Tri-Range http://www.lucidimagination.com/blog/2009/05/13/exploring-lucene-and-solrs-trierange-capabilities/

I know that these two solutions are different, do I need comments that should be given higher priority? Does the second solution fit my problem? And if more solutions come out for my spell check.

Thanks in advance

Answer 1

Indexing usually makes queries slow. If you have fast disks, indexing will use 100% of the CPU, otherwise it will use 100% of the disk bandwidth. In any case, the requests will be slow.

For this, the master / slave configuration is the standard solution. Slave servers are for search queries. The only time they slow down after replication is when new search engines are created with new caches.

The master / slave configuration may not make indexing much faster, but it will avoid slow query performance. There has been work on multithreading indexing, so you can try several indexing tasks at once. This will not help if the bottleneck is the IO drive, only if it uses 100% of one processor.

Trie fields are great for range queries. I doubt that they will greatly influence the speed of indexing.

Finally, you can adjust spelling options. Spelling can be a lot of work, and you can get good results with different, less expensive options.