Rails中Delayed Job执行两次问题排查与修复求助

Hey there! Let's break down why your SimpleJob is running twice and how to fix it. Based on your setup, here are the most likely causes and solutions:

1. Initializer Script Gets Loaded Multiple Times #

You mentioned adding SimpleJob.schedule! in config/initializers/job_initializers.rb. Rails can re-run initializers in certain scenarios—like code reloading in development mode, or when using a server that spins up multiple worker processes (Puma/Unicorn). Each reload/process launch triggers the initializer again, scheduling the job twice.

Fix:
Add a guard clause to ensure the job is only scheduled once:

# config/initializers/job_initializers.rb
# Option 1: Check if the job already exists in Delayed Job
unless Delayed::Job.exists?(handler: /SimpleJob/)
  SimpleJob.schedule!
end

# Option 2: Use a global flag to prevent repeats
unless defined?($simple_job_already_scheduled)
  $simple_job_already_scheduled = true
  SimpleJob.schedule!
end

2. Multiple Server Workers Trigger Scheduling #

If your app server runs multiple worker processes (common in production or even development with Puma's default settings), each process will execute the initializer independently—resulting in duplicate job schedules.

Fix:
Move the scheduling logic out of initializers and into a one-time rake task. Create lib/tasks/schedule_jobs.rake:

namespace :jobs do
  desc "Schedule one-time recurring jobs"
  task schedule: :environment do
    SimpleJob.schedule!
  end
end

Run it manually after starting your server: rails jobs:schedule, or add it to your deployment script (like Capistrano) to run once per deploy.

3. Delayed Job's Retry Mechanism Is Triggering #

If your job hits a temporary error (e.g., a brief database connection blip) during execution, Delayed Job defaults to retrying the task. If the error resolves quickly, you'll see two execution logs and two database records.

Fix:
Check your logs for retry messages, then adjust the job's retry policy:

class SimpleJob < ApplicationJob
  queue_as :default
  # Disable retries entirely if you don't need them
  retry_on StandardError, attempts: 0
  # Or set longer wait times to avoid immediate repeats
  # retry_on StandardError, wait: 5.minutes, attempts: 3

  def perform
    # Your job logic here
  end
end

4. Flawed Scheduling Logic in schedule! #

If your SimpleJob.schedule! method doesn't check for existing pending jobs, every call will create a new task—leading to duplicates.

Fix:
Update the schedule! method to only create a job if none are pending:

class SimpleJob < ApplicationJob
  def self.schedule!
    # Skip if there's already a pending SimpleJob
    unless Delayed::Job.where(handler: /SimpleJob/).exists?
      set(wait: 1.minute).perform_later # Adjust wait time as needed
    end
  end
end

5. Development Mode Code Reloading #

In development, Rails auto-reloads classes when you modify code, which can re-run initializers and re-schedule the job.

Fix:
Avoid scheduling jobs via initializers in development. Instead, run SimpleJob.schedule! manually in the Rails console when you need it, or use the rake task mentioned earlier.

内容的提问来源于stack exchange，提问作者onder