Challenge #5a: Single-Node Kafka-Style Log: what is stats.by_f.crash?

1

what is stats.by_f.crash?

:stats {:valid? true,
         :count 17291,
         :ok-count 17288,
         :fail-count 0,
         :info-count 3,
         :by-f {:
                :crash {:valid? false,
                        :count 3,
                        :ok-count 0,
                        :fail-count 0,
                        :info-count 3

What is crash.info-count?

What are those :crash invokes?

INFO [2023-03-17 13:04:49,673] jepsen worker 1 - jepsen.util 3	:invoke	:crash	nil
INFO [2023-03-17 13:04:49,674] jepsen worker 1 - jepsen.util 3	:info	:crash	nil
INFO [2023-03-17 13:04:49,674] jepsen worker 0 - jepsen.util 0	:invoke	:crash	nil
INFO [2023-03-17 13:04:49,674] jepsen worker 0 - jepsen.util 0	:info	:crash	nil
INFO [2023-03-17 13:04:49,676] jepsen worker 1 - jepsen.util 5	:invoke	:assign
2

Are you processes exiting? I’ve seen that when a process panics before.

The Maelstrom protocol docs are a bit more ambiguous:

Crash - Indicates that some kind of general, indefinite error occurred. Use this as a catch-all for errors you can’t otherwise categorize, or as a starting point for your error handler: it’s safe to return internal-error for every problem by default, then add special cases for more specific errors later.

3

No, my processes are not crashing and especially the overall maelstrom results are valid true.

Those are invokes from maelstrom workload. The same kind of strange as :assign invokes. Who knows whatever they are.

The doc you have mentioned is for the error codes that the nodes can return to the maelstrom dispatcher.

I suppose the original workload wants to test durability and maybe tries to crash the nodes. But what is the assign invoke I have no idea. On the other hand the system is not complaining that there is no crash handler in my code.