NameError: uninitialized constant UniqueTextWorker

[kshnurov]

I have a typical worker:

class UniqueTextWorker
  include Sidekiq::Worker
  sidekiq_options :queue => :background

  def perform(place_id)
...

When I'm trying to run some workers (seriosly - some place_ids work, some not), I'm getting an error: "NameError: uninitialized constant UniqueTextWorker".
Code is totally fine and successfully working outside of sidekiq. Backtrace is useless. What's wrong?

activesupport-3.2.11/lib/active_support/inflector/methods.rb:230→ block in constantize
activesupport-3.2.11/lib/active_support/inflector/methods.rb:229→ each
activesupport-3.2.11/lib/active_support/inflector/methods.rb:229→ constantize
activesupport-3.2.11/lib/active_support/core_ext/string/inflections.rb:54→ constantize
sidekiq-2.12.0/lib/sidekiq/processor.rb:43→ process
celluloid-0.14.0/lib/celluloid/calls.rb:25→ public_send
celluloid-0.14.0/lib/celluloid/calls.rb:25→ dispatch
celluloid-0.14.0/lib/celluloid/calls.rb:125→ dispatch
celluloid-0.14.0/lib/celluloid/actor.rb:328→ block in handle_message
celluloid-0.14.0/lib/celluloid/tasks.rb:42→ block in initialize
celluloid-0.14.0/lib/celluloid/tasks/task_thread.rb:20→ block in create
celluloid-0.14.0/lib/celluloid/internal_pool.rb:59→ call
celluloid-0.14.0/lib/celluloid/internal_pool.rb:59→ block in create

[darkside]

Assuming you're using Rails, this could be related to autoload, I've seen this happen. Is your worker sitting under lib/ ?

[kshnurov]

No, it's in app/workers and here's config:

config.autoload_paths += %W(#{config.root}/lib)
config.autoload_paths += Dir["#{config.root}/lib/**/"]
config.eager_load_paths += ["#{config.root}/lib"]
config.eager_load_paths += ["#{config.root}/lib/**/"]

[darkside]

Yeah mine are sitting under app/workers as well, and I have on my application.rb:

config.autoload_paths += %W(#{config.root}/app/workers)

😕

[kshnurov]

Yeah, now it works. Thank you!

[kshnurov]

Now I'm getting that error only with big queues - first 25 workers starting ok, all the next failing.
I've tried to add sidekiq-limit_fetch gem and limit all queues to 20 workers as a workaround - it didn't help. All workers, except first 25, failing immediately after start.

[kshnurov]

Even when sidekiq is paused (bundle exec sidekiqctl quiet) - all workers fail immediately after adding to queue.

[darkside]

Try using config.eager_load_paths += instead of autoload. See https://github.com/mperham/sidekiq/wiki/FAQ (Why doesn't Sidekiq autoload my Rails application code?)

[kshnurov]

I'm already using both.

[darkside]

Only other thing I can think about is there could be something else breaking and preventing your class from being loaded. I never had your second issue, it either always found out the class or never found the class. 😐

[dmeremyanin]

Try to explicitly require your worker code before the application is defined.
In config/application.rb:

require './app/workers/unique_text_worker'

module YourApplication
  class Application < Rails::Application
    # ...

[kshnurov]

dimko, it didn't help. I've updated sidekiq from 2.14.0 to 2.14.4 - still nothing.

[TheCorp]

Another option is to try initializing (using in any way) the class in the sidekiq initializer. That should make sure its loaded into memory before any of the workers get to using it.

[Bodacious]

I'm having a similar issue just now...

All of my mailer jobs are failing because the class can't be found: undefined class/module MemberMailer

All mailers are in app/mailers/*

Rails version 3.2.14
Sidekiq version 2.13.0

We go live tomorrow! Any advice would help right now

[kshnurov]

In my case I had old Sidekiq process that was not terminated and was hanging in "terminating" status for a week. Try to kill and restart all sidekiq processes.

[Bodacious]

Hey @kshnurov thanks for the feedback.

This was actually a developer error, I'd removed the Sidekiq.configure_client block from the initializer 😟

[ghost]

If this is happening locally, make sure that config.cache_classes = true in development.rb. The docs say that config.eager_load_paths only works if class caching is enabled (this sucks for development though).

http://guides.rubyonrails.org/v3.2.13/configuring.html#rails-general-configuration

config.eager_load_paths accepts an array of paths from which Rails will eager load on boot if cache classes is enabled. Defaults to every folder in the app directory of the application.

This worked for me in development with Rails 3.2.13, Sidekiq 2.12.0.

[mperham]

Sidekiq always eager loads due to thread safety. It does not support development mode and ignores cache_classes.

On Aug 1, 2013, at 18:48, Christopher Chow notifications@github.com wrote:

If this is happening locally, make sure that config.cache_classes = true. The docs say that config.eager_load_paths only works if class caching is enabled (this sucks for development though).

http://guides.rubyonrails.org/v3.2.13/configuring.html#rails-general-configuration

config.eager_load_paths accepts an array of paths from which Rails will eager load on boot if cache classes is enabled. Defaults to every folder in the app directory of the application.

—
Reply to this email directly or view it on GitHub.

[vkeziah]

I am getting NameError (uninitialized constant Api::V1::EmailController::ResponseWorker) when I try to call workers from my controller , I added autoload paths still getting this error , can anyone please help me

[bcobb]

@vkeziah Some information that would be helpful in getting to the bottom of your problem:

1. What version of Ruby/Rails/Sidekiq are you using?
2. Does this error occur in every environment? Only development? Only production?
3. What is the path to your worker?
4. What are the first few lines of the worker file - at least until include Sidekiq::Worker?
5. What is the line of code in the controller that sends a job to the worker?

[vkeziah]

1. I am using ruby 2.1.5, rails - 4.1.x , sidekiq -3.1.x
2. I check in only in development
3. my workers are reside in app/workers folder
4. require 'sidekiq'
   require 'logger'

class ResponseWorker
include Sidekiq::Worker

1. ResponseWorker.perform_async(id)

Please let me know if you nee any more information

[seuros]

@vkeziah : upgrade to sidekiq 3.3.x and try again.

[vkeziah]

yep it works ... thanks

However I am facing wiht another problem .. when I try to find the records from the db I am getting error stack level too deep , not sure where the error is but I am suspecting that its with db connection and thought that it was unable to establish the connection with db to find the records.

Can you please let me know how to log the exact errors and way to work with database from workers.

[mperham]

Please don't hijack old issues; open your own issue. Usage questions belong on the mailing list.