Fork me on GitHub

Analyzing the problem I sometimes have with my job using the datomic plugin read-log which for some reason is processing again :t, I came across this exception in the log.

11:20:56.341 INFO  f.onyx.builder.observable - read-log TX ->  743070
11:21:50.475 ERROR f.onyx.builder.core - Lifecycle:  {:lifecycle/task :all, :lifecycle/calls :formiguinhas-service.onyx.builder.core/calls, :lifecycle/doc Chamado em todo lifecycle, inclui o tratamento de exception a todas as tasks}
11:21:50.475 ERROR f.onyx.builder.core - Lifecycle Phase:  :lifecycle/offer-heartbeats
11:21:50.988 ERROR f.onyx.builder.core - Event:  {:onyx.core/outbox-ch #object[clojure.core.async.impl.channels.ManyToManyChannel 0x7ded4a59 clojure.core.async.impl.channels.ManyToManyChannel@7ded4a59], :onyx.core/batch [], :onyx.core/log-prefix Job 78563630-7273-38fd-5a2e-817d55e2d596 {:job-id #uuid "78563630-7273-38fd-5a2e-817d55e2d596", :job-hash "35a180648c4f9d0cd356be934279abc364279ccc15c86dce49b8a8186fa15"} -
11:21:51.077 ERROR f.onyx.builder.core - ERRO-HANDLER-ALL Alguma exception diferente da Amazons3Exception entao
                  Returning false, indicating that this task should NOT restart. e: 
 com.amazonaws.SdkClientException: Unable to execute HTTP request: Socket is closed
11:21:52.161 INFO  f.onyx.builder.observable - read-log TX ->  727933
In the case of this exception I was returning false, but I saw in the documentation that I should return: kill if I want to kill the job. If I return false how is it today which behavior takes the job? In case I want to restart, and using the resume point how to protect me so that it does not run again :t that has already been processed? Because if you see in the log, the job is processing :t 743070 but after the exception it back to 727933. Tks for help