reject! cause hutch worker to shutdown

[gottfrois]

Any idea why reject! kill the worker? Am I missing something?

module Consumers
  class Test
    include ::Hutch::Consumer

    consume 'test'

    def process(response)
      reject!
    end
  end
end

2015-10-15T14:47:27Z 34520 INFO -- hutch booted with pid 34520
2015-10-15T14:47:27Z 34520 INFO -- found rails project (.), booting app
2015-10-15T14:47:43Z 34520 INFO -- connecting to rabbitmq (amqp://[email protected]:5672/immvnaiq)
2015-10-15T14:47:43Z 34520 INFO -- connected to RabbitMQ at 127.0.0.1 as immvnaiq
2015-10-15T14:47:43Z 34520 INFO -- opening rabbitmq channel with pool size 1
2015-10-15T14:47:43Z 34520 INFO -- using topic exchange 'textmaster'
2015-10-15T14:47:43Z 34520 INFO -- HTTP API use is disabled
2015-10-15T14:47:43Z 34520 INFO -- tracing is disabled
2015-10-15T14:47:43Z 34520 INFO -- setting up queues
2015-10-15T14:48:35Z 34520 INFO -- message(278fc81e-f7a8-4655-b398-faed1a049cca): routing key: test, consumer: Consumers::Test, payload: {}
2015-10-15T14:48:35Z 34520 INFO -- hutch shut down gracefully

Rabbitmq logs:

=INFO REPORT==== 15-Oct-2015::16:51:18 ===
accepting AMQP connection <0.14570.0> (127.0.0.1:56974 -> 127.0.0.1:5672)

=ERROR REPORT==== 15-Oct-2015::16:51:18 ===
Channel error on connection <0.14570.0> (127.0.0.1:56974 -> 127.0.0.1:5672, vhost: 'immvnaiq', user: 'immvnaiq'), channel 1:
{amqp_error,precondition_failed,"unknown delivery tag 1",'basic.ack'}

=WARNING REPORT==== 15-Oct-2015::16:51:18 ===
closing AMQP connection <0.14570.0> (127.0.0.1:56974 -> 127.0.0.1:5672):
connection_closed_abruptly

[teodor-pripoae]

I think it is related to #165. You are sending the reject and then the worker sends ack for the same message, causing the connection to drop.

[michaelklishin]

Yes, the issue here is double acking (or reject-then-ack), which results in a channel-level protocol exception. #165 is indeed aims to solve this problem. See also #177, which has a decent chance at being merged.