2 ms·
I spent time frequently mucking with the elasticsearch codebase from 5.x through till 6.4 strictly from durability and resiliency PoV [0]. The improvements it i
by ignoramous 7y ago
I spent time frequently mucking with the elasticsearch codebase from 5.x through till 6.4 strictly from durability and resiliency PoV [0]. The improvements it is seeing in 7.x are tremendous. It still had the occasional tendency to go hay-wire and show behaviour that simply didn't make sense and couldn't be recovered from, but I'm confident 7.x will push it to making it even more resilient than 5.x+ [1].
The sheer number of configurable options for the cluster bringup, index setup, plugable components turns it into a very hard to tame beast. There is an opportunity here somewhere for someone to make Elasticsearch on Rails (convention over configuration). With 7.x, I feel the time is just abt right to try something like that.
What excites me most abt 7.x is the overhaul of underlying distributed engine that's been going on for over 2yrs [2] that will enable a whole slew of super nice features [3] that'd put it on par with state-of-the-art distributed databases in terms of not only features but resiliency.
[0] https://www.elastic.co/guide/en/elasticsearch/resiliency/current/index.html https://www.elastic.co/guide/en/elasticsearch/resiliency/cur...
[1] https://www.microsoft.com/en-us/research/publication/pacifica-replication-in-log-based-distributed-storage-systems/ https://www.microsoft.com/en-us/research/publication/pacific...
[2] https://github.com/elastic/elasticsearch/issues/10708 https://github.com/elastic/elasticsearch/issues/10708
[3] https://www.elastic.co/blog/elasticsearch-sequence-ids-6-0 https://www.elastic.co/blog/elasticsearch-sequence-ids-6-0