Follow up - 6839f040e0 - Fixed bug: Failed jobs get executed more times than configured.

This commit is contained in:
Thorsten Eckel 2019-04-17 14:24:52 +02:00
parent 093d8e4722
commit b6f277c97b
2 changed files with 27 additions and 1 deletions

View file

@ -0,0 +1,26 @@
# In previous versions of Zammad we used Delayed::Job exclusively
# for performing background jobs. Delayed::Job was therefore in
# charge of scheduling, retrying and executing background jobs.
# After the (partly) migration to Rails ActiveJob this has changed.
# Now ActiveJob is in charge of scheduling and retrying jobs
# while Delayed::Job is still in charge of executing the jobs (assigned to it).
# That leads to an issue where Delayed::Job now falls back to the default
# of 25 retries for a failed job.
# This is not wanted since retries are handled by ActiveJob.
# Therefore the JobWrapper (the class/handler that gets queued) has to define
# max_attempts to be only one. A failing ActiveJob will now be retried as
# often as configured by ActiveJob and then an exception will be raised by
# ActiveJob with the last error message. This message will be rescued by
# Delayed::Job which will see that there are no more attempts wanted and
# will record and handle it as a failed job.
module ActiveJob
module QueueAdapters
class DelayedJobAdapter
class JobWrapper
def max_attempts
1
end
end
end
end
end

View file

@ -270,7 +270,7 @@ class EmailDeliverTest < ActiveSupport::TestCase
assert_raises(RuntimeError) do
Scheduler.worker(true)
end
assert(Delayed::Job.where(attempts: 4).exists?)
assert(Delayed::Job.none?)
ticket1.reload
article2_lookup = Ticket::Article.find(article2.id)