The increase is data and traffic accompanied with the feature set offered by Solr has led to an increased usage of Solr in recent times. As a lot of these applications are critical, a reliable disaster recovery mechanism has become essential.
While Solr already offers a way to accomplish disaster recovery using cross-dc replication out of the box, there are other ways to accomplish the same. Copying data is the obvious part of this system but a lot of supporting features or systems like monitoring, self-defense mechanism etc. are needed for a complete cross-dc solution.
A simple solution, like the one that Solr offers out of the box works 90% of the times, but the lack of supporting features, monitoring tools, and also self-defense mechanism make it a difficult choice in a lot of places. Using such a system might catch you off-guard and leave you with a false sense of DR availability.
During this talk, I would like to talk about a few approaches of achieving cross-dc availability in Solr, something that I have tried and used over the years and highlight the pros and cons of each of them. I would also talk about all the satellite features and applications that are needed to ensure consistency of these clusters in addition to the self-healing, and self-defense mechanisms that are needed to reliably run and use the DR clusters.
At the end of this talk, attendees would have a much better understanding of achieving cross-dc for Solr and their reliability levels. They would also have learnt about the supporting systems that are needed to have a solid cross-dc story in Solr, one that scales and works reliably.