Lots of organizations search for fault-oriented solutions at a high level and this depends on the different parts of the infrastructure. The solr support is known to be one of the most mature solutions for all types of search and this is being available for a decade. This search solution is based on conventional master-slave deployment and this technology is in use since the year 2006. This is the reason for which lots of organizations are trying to migrate the process to Solr Cloud from Solr Master-slave. The process of this migration can be mentioned below.
- Hardware: The first step that one needs to plan at the time of migrating from the old master-slave environment is nothing but the hardware. Ensure to maintain current production at the time of migration. Design the infrastructure for the new search setup to serve successfully in the coming time. Indexing of the data to all nodes needs to be done. Plan for calculative replicas keeping in mind available storage requirement. Ensure the ZooKeeper ensemble is working. SolrCloud makes use of ZooKeeper to turn it fully configurable.
- Solar Setup: Setting of SolrCloud will be different from Solr master-slave architecture. Find Apache ZooKeeper ensemble. SolrCloud takes advantage of ZooKeeper for collections, collections state, configuration, tracking of nodes, and many others. In case it’s not available indexing operation will experience failure. Professionals of solr consulting will also help to test this solr cloud whether it is perfect or not. For setting up highly available yet fault-tolerant ZooKeeper ensemble there is need of minimum 3 instances. Which allows a single instance of the ZooKeeper to move down and get the ensemble in running mode.
- Migration of configuration: The next stage will be migration from the Solr master-slave to SolrCloud for preparing configuration files. There will be a minimum of two files – the first one will be schema.xml and the second one will be solrconfig.xml and it’s important to take care of it. An important point that needs to be noted here is that additional files requirement may be there or may not be and need to be removed, based on the nature or structure of the configuration changes.
- Migration of Data: In maximum cases, when one migrates data from the old Solr version to the new one, for example, upgrading of SolrCloud to major versions where one can avoid data re-indexing. But at time of migrating Solr master-slave environment to SolrCloud things will be tricky and might not worth it.
There are few more things we wish you to consider while migrating to Solr Cloud. At the time of moving to Solr 7, you will come across changes which will be backward incompatible. Let’s take an example here: Lucene no more supports the index-time boosting. Well that means that one needs to look for a different way for handling the boosts in case the user is still making use of them – function queries as well as query time boosts are one way to move but that will need some minimum amount of testing.
In case things look complicated for you then seeking the assistance of the solr consulting experts will make it easy and simple for you to implement.