With the release of Solr 7.0 and introduction of new replica types, in addition to the defa?ult NRT type the question appeared – can we control the queries and where they are executed? Can we tell Solr to execute the queries only on the PULL replicas or give TLOG replicas a priority? Let’s check that out.Read more
One of the questions I tend to get is what happens with SolrCloud cluster when ZooKeeper fails. Of course we are not talking about a single ZooKeeper instance failure, but the whole ensemble not being accessible and so the quorum not present. Because the answer to this question is very easy to verify i decided to make a simple blog post to show what happens when ZooKeeper fails.
What is the most important change in 4.x version of Apache Solr? I think there are many of them but Solr Cloud is definitely something that changed a lot in Solr architecture. Until now, bigger installations suffered from single point of failure (SPOF) – there was only the one master server and when this server was going down, the whole cluster lose the ability to receive new data. Of course you could go for multiple masters, where a single master was responsible for indexing some part of the data, but still, there was a SPOF present in your deployment. Even if everything worked, due to commit interval and the fact that slave instances checked the presence of new data periodically, the solution was far from ideal – the new data in the cluster appeared minutes after commit.