With the updates to how we send validation job errors to specific slack channels, we now will run multiple batches of jobs with the ExternalJobScheduler, which uses Parallel to schedule jobs. This means we will need to reconnect to the database before each job run in the ExternalJobScheduler too. (See the ParallelJobScheduler for an ex of having to reconnect to the Postgres database for each parallel job run)
Note: I bumped the ruby version but have not published the gem yet. Will publish once the PR is merged
Recently our HIValidations job has been failing with connection issues: https://jenkins.zp.int/view/Validations/job/HIValidation/2349/console
Through some investigation, it looks like the Parallel gem has an interesting bug with Postgresql where it will disconnect from the database after each job run. http://ruby.zigzo.com/2012/01/29/the-parallel-gem-and-postgresql-oh-and-rails/
With the updates to how we send validation job errors to specific slack channels, we now will run multiple batches of jobs with the ExternalJobScheduler, which uses Parallel to schedule jobs. This means we will need to reconnect to the database before each job run in the ExternalJobScheduler too. (See the ParallelJobScheduler for an ex of having to reconnect to the Postgres database for each parallel job run)
Note: I bumped the ruby version but have not published the gem yet. Will publish once the PR is merged