Fatal error (machine stack overflow) on job retry

[romiras]

trafficstars

Ruby version: 2.7.x (latest) Rails version: 6.0.x (latest) Sidekiq Pro version: 5.5.x (latest)

I cannot share sidekiq.yml in public, though I can send it in private, on demand. It includes these requires:

require "honeybadger/plugins/sidekiq"
require 'sidekiq/job_logger'
require "sidekiq/throttled"
require 'sidekiq'
require("sidekiq-unique-jobs")
require 'sidekiq/worker_killer'
require 'sidekiq/middlewares/memlog_middleware'
require 'sidekiq/middlewares/mysql_connection_middleware'
require 'sidekiq/middlewares/profile_middleware'
require 'sidekiq/max_rss_calculator'

Exception trace: Nested Exceptions https://github.com/sidekiq/sidekiq/blob/fe04aa110ea3144bdd12a8bc80dadfc5c560b06e//lib/sidekiq.rb#L124 https://github.com/sidekiq/sidekiq/blob/fe04aa110ea3144bdd12a8bc80dadfc5c560b06e//lib/sidekiq/processor.rb#L266

Error message:

fatal: machine stack overflow in critical region

Error trace in Honeybadger:

{
  "context" => "Job raised exception",
  "job" => {
    "retry" => true,
...
    "jid" => "f5b408ccd1efc34d2be7cca7",
    "created_at" => 1711225525.286979,
    "enqueued_at" => 1711225525.2877622
  }
}

This error happens in different time. I didn't found related in later versions in change log.

[IvanTakarlikov-st]

@mperham We're experiencing the same issue with Ruby 2.7.6, Rails 6.1 and sidekiq-pro 7.2 and sidekiq-cron 1.12.0. Could you please advice where might be an issue, or any details that might help with investigation?

[mperham]

That’s a bug in Ruby (or possibly a native extension). Upgrade them.