ucsdlib / damspas-rd

A Digital Collections application based on Hyrax
MIT License
3 stars 2 forks source link

Sidekiq optimizations #152

Open mcritchlow opened 6 years ago

mcritchlow commented 6 years ago

Descriptive summary

I'm placing this here so we don't forget..

https://www.mikeperham.com/2018/04/25/taming-rails-memory-bloat/

Basically, once we start using Sidekiq for DAMS5, we should experiment with doing both of these things. The differences in memory usage are... startling, and if we were paying cloud costs this would be a very expensive thing to not do

  1. Set: MALLOC_ARENA_MAX = 2
  2. Compile Ruby with --with-jemalloc

Rationale

More info: https://bugs.ruby-lang.org/issues/14718