Solr/Lucene Meetup January 2018

Session #1: Leveraging PULL replicas in Solr 7 for improving performance

Abstract: On a multi cluster NRT replicas setup in Solr 6, where documents are updated real-time, there were challenges in terms of N/W bandwidth utilisation on leader nodes, and the indexing response times. With the introduction of TLOG and PULL replica types in SolrCloud in Solr 7, we have leveraged them to improve the N/W utilisation and to reduce the indexing RTs. With the introduction of metrics reporters for different interfaces, the dependency on 3rd party tools for monitoring Solr is also removed.

Speaker Bio: Samuel Tatipamula is an SSE in Search Team at Myntra. He has been working on Myntra's search backend systems since past 1 year. He has worked on scaling and performance optimisation of Solr.

Session #2: Introduction to autoscaling

Abstract:
In Solr 7.0, there was a new feature introduced on autoscaling Solr based on predefined policies. In the talk, I wish to provide a brief overview of the capabilities offered, and follow it up with a brief demo, and the road ahead.

Speaker Bio: Apoorv Bhawsar, Search Engineer, Unbxd

Session #3: Upcoming Features in Solr

Abstract:
There are some features which are going to be added to Solr in future releases, related to SchemaTraining, and TimeRoutedPartitioning of collections.. In the talk, i wish to provide a brief overview of the features and want to follow it up with a small demo to cover some use-cases. The features include:-

1. Time Routed Collection Partitions (SOLR-11722)
2. How to Train Your Schema (SOLR-11741)

Speaker Bio: Abhishek Kumar Singh, Search Engineer, Unbxd