What is the difference between Solr Replication and Solr Cloud?

I support the rails project, which contains the rails application and an additional instance with Solr.

My environment: rails 3.2.1, ruby ​​2.1.2, sunny spotting 2.1.0, Solr 4.1.6.

Problem:

The Clould provider is unstable. I cannot use another cloud provider - this is consumer demand. Often, the search server continues to service, and the web application stops working in production. So, I’m thinking about how to make 2 identical search engines instead of one in order to make the system more stable: if one server is disconnected, the others will continue to work.

The way out of the problem:

Question:

  • What to do and why?
  • Is SolrCloud the best solution?
  • What are the PROC and CONS of these options?
  • When is it better to combine SolrCloud?
  • When is it better to combine Solr replication?
+4
source share
1 answer

You can find comparisons for the Solr Master Slave Heirarchy Vs SolrCloud in the forums, but this is my understanding / experience from them, so use it easily.

  • What to do and why? It depends entirely on the requirements of your business. I would say that if your application simply reads data from SOLR and requires high availability when reading data from SOLR, then a simple single wizard is more than enough for many subordinate hierarchies. But if you are looking for high availability when writing to SOLR, then SolrCloud is the right option.
  • Is SolrCloud the best solution? There is nothing better, because SolrCloud needs a good infrastructure and needs to monitor the availability of ZooKeepers, as well as the health of nodes, high-performance disk for better replication speed, etc. But other than that, we don’t have to worry about data consistency between nodes, as this will be done by SolrCloud
  • What are the PROC and CONS of these options? As described above.
  • When is it better to combine SolrCloud? When you need high availability in Solr Writes, as well as reading, we should go for SolrCloud. In addition, if you cannot allow larger machines to have one single node, then you can split your index into fragments and store it in smaller configuration machines.
  • When is it better to choose Solr Replication? When your application does not write SOLR in real time, replication is sufficient and there is no need to complicate the work with SolrCloud. In addition, its relatively simple to configure Master Slave than SolrCloud
+5
source

All Articles