lock remains after job not properly finish

[mathieujobin]

Describe the bug
We have had this issue with one specific job, where the unique lock would remain in redis after the job has crashed or did not finish properly

is this the intended behavior?

Expected behavior
the lock should be cleared after the job finish, regardless of the status

Current behavior
new jobs gets schedule and are discarded indefinitely

Worker class

class MyWorker
  include Sidekiq::Worker

  sidekiq_options :retry => false, :backtrace => true, :unique => :while_executing

  def perform(args); raise("I made a mistake"); end
end

Additional context
I noticed you are talking there was a 30 minutes timeout in the 5.x series which got removed.
but in your readme you seems to be recommending against using lock_expiration/lock_timeout
I am also unsure which one would fit best my use case.

thank you for your help

[mathieujobin]

oh, this is exact duplicate of #332

[mathieujobin]

reopening, I'd appreciate some insight as to know if lock_expiration or lock_timeout could help ?

[mhenrixon]

Lock expiration definitely will help. The timeout is just how long we wait for a lock to be achieved

[mathieujobin]

thanks