GoogleCloudPlatform/ubbagent

message dequeue panic

Closed this issue · 0 comments

I encountered an error when I send a message one but I see multiple invocations of the same message in retry.go
the message gets dequeued after the first success, when it attempts to do so in the second attempt the dequeue fails as the message is not in the queue anymore as it was dequeued in the first attempt.
This happens in case of service control endpoint and it is not an expected behaviour

To make things worse ubbagent panics resulting in the system to crash
I believe when this happens it should only log the error and move on

So we have 2 problems at hand that we must discuss:

  1. why should the dequeue panic when we can just log the error
  2. why would it try to send it twice in the first place when it already had a success sending it in the service control