sunspot-queue
sunspot-queue copied to clipboard
gaffneyc
Background search indexing using existing worker systems
Hey all, We're seeing quite a few jobs in our queue for objects that were destroyed — can we add a check for `destroyed?` (well, if it's an ActiveRecord instance)...
Currently sunspot-queue will fail to retrieve active records if default scopes filter them out. This PR unscopes the default scopes first so that they're always found.
Would love to have support for ActiveJob. I'd be willing to submit a PR or create a separate gem for it.
Sidekiq uses threads to handle multiple workers, and the without_proxy helper is not threadsafe because it modifies Sunspot.session. The Sunspot::Queue::Sidekiq job workers can be modified to work safely like this:...
Solr processes updates most efficiently when it's receiving documents in batches at a relatively low concurrency. Highly parallel single document updates are probably one of the easiest ways to starve...
This allows the instance to determine if it should be enqueued for index adds/removals based on things like certain states. So an order for example would not be enqueued at...
A couple forks have added delay support. I don't fully understand the reasoning or what is causing the need for it. My best assumption is that the worker system is...
Temporary work around: ``` ruby Sunspot::Queue::Resque::IndexJob.class_eval do def self.queue "low" end end Sunspot::Queue::Resque::RemovalJob.class_eval do def self.queue "low" end end ```
On a project that was using IndexQueue we would occasionally see certain indexing jobs completely lock up. Timeouts would keep the workers from getting completely locked up.