Ivor Wanders

Was about to file a PR and I saw this issue and related #119. Just quickly chiming in, since this is really simple since my contributions in PR's #116, #108...

I think this is the same error as I encountered and fixed in #108.

A bit late, but I wrote some explanation on #120. I hope that helps.

Thanks for taking a look at this @ajs124, appreciated! > Can you maybe write a test for this? I think this would probably fit into `t/Hydra/Controller/API/checks.t`. Certainly, I've added one...

I do sometimes see the following warnings ``` ( STDERR ) job 87 (in cleanup) {UNKNOWN}: DBI Exception: DBD::Pg::st DESTROY failed: no connection to the server [for Statement "INSERT INTO...

> Sockets for PGHOST should also work. I switched it to use tcp to allow easy use of the [postgres](https://hub.docker.com/_/postgres) docker container. Exposing a unix domain socket should work as...

> You did what? This diff makes it seem like it was using TCP before and after the change. I am fine with that, but I would be wary of...

I think this captures it. We can do logging to `stderr` if we are writing the traces to `stdout`. That way we can still get info if piping it into...

Yeah, we should be able to do that by moving parts of [this](https://github.com/iwanders/scalopus/blob/master/scalopus_tracing/src/native/native_trace_source.cpp#L70) function to another function that `yields` a vector of jsons from the internal collection container (`recorded_data_`) in...

> Following example illustrates such crash scenario: https://github.com/aurzenligl/study/blob/master/ros-timer/src/race.cpp Yikes, that's definitely not good. Thanks for a minimal example to reproduce this failure. > I assume proper/expected/consistent behavior is to block....